Indoor farms, also known as vertical farms, are popular among agricultural researchers and are expanding across the agricultural industry. Some benefits they have over outdoor farms include:

• Year-round production of food crops
• Less water and land requirements
• Not needing pesticides
• Reducing carbon emissions from shipping
• Reducing food waste

Additionally, some studies indicate that indoor farms produce more nutritious food for urban communities. 

However, these farms are often inaccessible to birds, bees, and other natural pollinators, leaving the pollination process to humans. The tedious process must be completed by hand for each flower to ensure the indoor crop flourishes.

Ai-Ping Hu, a principal research engineer at the Georgia Tech Research Institute (GTRI), has spent years exploring methods to efficiently pollinate flowering plants and food crops in indoor farms to find a way to efficiently pollinate flower plants and food crops in indoor farms.

Hu, Assistant Professor Shreyas Kousik of the George W. Woodruff School of Mechanical Engineering, and a rotating group of student interns have developed a robot prototype that may be up to the task.

The robot can efficiently pollinate plants that have both male and female reproductive parts. These plants only require pollen to be transferred from one part to the other rather than externally from another flower.

Natural pollinators perform this task outdoors, but Hu said indoor farmers often use a paintbrush or electric tootbrush to ensure these flowers are pollinated. 

Knowing the Pose

An early challenge the research team addressed was teaching the robot to identify the “pose” of each flower. Pose refers to a flower’s orientation, shape, and symmetry. Knowing these details ensures precise delivery of the pollen to maximize reproductive success. 

“It’s crucial to know exactly which way the flowers are facing,” Hu said.

“You want to approach the flower from the front because that’s where all the biological structures are. Knowing the pose tells you where the stem is. Our device grasps the stem and shakes it to dislodge the pollen.

“Every flower is going to have its own pose, and you need to know what that is within at least 10 degrees.”

Computer Vision Breakthrough

Harsh Muriki is a robotics master’s student at Georgia Tech’s School of Interactive Computing, who used computer vision to solve the pose problem while interning for Hu and GTRI.

Muriki attached a camera to a FarmBot to capture images of strawberry plants from dozens of angles in a small garden in front of Georgia Tech’s Food Processing Technology Building. The FarmBot is an XYZ-axis robot that waters and sprays pesticides on outdoor gardens, though it is not capable of pollination.

“We reconstruct the images of the flower into a 3D model and use a technique that converts the 3D model into multiple 2D images with depth information,” Muriki said. “This enables us to send them to object detectors.”

Muriki said he used a real-time object detection system called YOLO (You Only Look Once) to classify objects. YOLO is known for identifying and classifying objects in a single pass.

Ved Sengupta, a computer engineering major who interned with Muriki, fine-tuned the algorithms that converted 3D images into 2D.

“This was a crucial part of making robot pollination possible,” Sengupta said. “There is a big gap between 3D and 2D image processing.

“There’s not a lot of data on the internet for 3D object detection, but there’s a ton for 2D. We were able to get great results from the converted images, and I think any sector of technology can take advantage of that.”

Sengupta, Muriki, and Hu co-authored a paper about their work that was accepted to the 2025 International Conference on Robotics and Automation (ICRA) in Atlanta.

Measuring Success

The pollination robot, built in Kousik’s Safe Robotics Lab, is now in the prototype phase. 

Hu said the robot can do more than pollinate. It can also analyze each flower to determine how well it was pollinated and whether the chances for reproduction are high.

“It has an additional capability of microscopic inspection,” Hu said. “It’s the first device we know of that provides visual feedback on how well a flower was pollinated.”

For more information about the robot, visit the Safe Robotics Lab project page.

In these contests, a handler directs a trained dog with whistle signals to guide a small group of sheep across a field and sometimes split the flock cleanly into two groups. But sheep do not always cooperate.

Researchers at the Georgia Institute of Technology studied how handler–dog teams manage these unpredictable flocks in sheepdog trials and found principles that extend beyond livestock herding.

In a study published in Science Advances as the cover feature, the researchers applied those insights to computer simulations showing how similar strategies could improve the control of robot swarms, autonomous vehicles, AI agents, and other networked systems where many machines must coordinate their actions despite uncertain conditions.

Group Movement Dynamics

“Birds, bugs, fish, sheep, and many other organisms move in groups because it benefits individuals, including protection from predators,” said Saad Bhamla, an associate professor in Georgia Tech’s School of Chemical and Biomolecular Engineering. “The puzzle is that the ‘group’ is not a single organism. It is built from many individuals, each making local, imperfect decisions.”

When a predator threatens a herd of sheep, individuals near the edge often move toward the center to reduce their own risk, Bhamla explained. “This is ‘selfish herd’ behavior,” he said. “Shepherds exploit that instinct using trained dogs.”

From examining hours of contest footage, the researchers found that controlling small groups of sheep can be harder than managing large ones. A larger group, with more sheep protected in the center, may behave more coherently than a small group as the animals constantly shift between two instincts: “follow the group” and “flee the dog.”

“That switching behavior makes the group unpredictable,” said Tuhin Chakrabortty, a former postdoctoral researcher in the Bhamla Lab who co-led the study.

Looking closely at how dogs and their handlers guide small groups, the researchers found that unpredictability in the flock’s behavior does not always make control harder. “Under the right conditions, that ‘noisy’ behavior might actually be a benefit,” Bhamla said.

Successful Sheep Herding

Sheepdog handlers categorize sheep by how strongly they respond to a dog’s threatening pressure. Some very responsive sheep might panic under too much pressure, while others might ignore mild pressure and require stronger positioning by the dog.

The researchers observed that successful control often followed a two-step pattern. First, the dog subtly influenced the sheep’s orientation while the animals were mostly standing still. Once the flock was aligned in the desired direction, the dog increased pressure to trigger movement. The timing of those actions was critical, because alignment within a small group could disappear quickly as individuals switched between instincts.

“In our simulations, increasing pressure makes the flock reach the desired orientation faster, but how long the flock stays aligned is set mainly by noise,” Chakrabortty said. “In essence, dogs can steer the direction, but they can’t hold that decision indefinitely, so timing matters.”

World foundation models (WFMs) enable physical AI systems to learn and operate within synthetic worlds created by generative artificial intelligence (genAI). For example, these models use predictive capabilities to generate up to 30 seconds of video that accurately reflects the real world.

The new framework, developed by a Georgia Tech researcher, enhances the processing speed of the neural networks that simulate these real-world environments from text, images, or video inputs.

The neural networks that make up the architectures of large language models like ChatGPT and visual models like Sora process contextual information using the “attention mechanism.”

Attention refers to a model’s ability to focus on the most relevant parts of input.

The Neighborhood Attention Extension (NATTEN) allows models that require GPUs or high-performance computing systems to process information and generate outputs more efficiently.

Processing speeds can increase by up to 2.6 times, said Ali Hassani, a Ph.D. student in the School of Interactive Computing and the creator of NATTEN. Hassani is advised by Associate Professor Humphrey Shi.

Hassani is also a research scientist at Nvidia, where he introduced NATTEN to Cosmos — a family of WFMs the company uses to train robots, autonomous vehicles, and other physical AI applications.

“You can map just about anything from a prompt or an image or any combination of frames from an existing video to predict future videos,” Hassani said. “Instead of generating words with an LLM, you’re generating a world.

“Unlike LLMs that generate a single token at a time, these models are compute-heavy. They generate many images — often hundreds of frames at a time — so the models put a lot of work on the GPU. NATTEN lets us decrease some of that work and proportionately accelerate the model.”

He also has cerebral palsy — and for the past four years, he has played a key role in improving one of the most exciting medical devices at Georgia Tech.

“My role here is as a participant in exoskeleton research studies,” Case explained. “When I come in, researchers hook me up to sensors that monitor my gait when I’m walking in the device, and then they get a whole lot of data based off that.”

Read more »

A team of Georgia Tech researchers wanted to ease that struggle, and robotic exoskeletons offered a promising path. Their findings point to a simple but powerful shift: exoskeletons that adapt to people, rather than forcing people to adapt to the machine. Using artificial intelligence (AI) to learn the rhythm of patients’ strides in real time, the team showed how these devices can reduce strain and increase efficiency. They also demonstrated how the technology can help restore confidence for stroke survivors. 

The Robot Finds the Rhythm

A robotic exoskeleton is a wearable device that helps people move with mechanical support. Traditional exoskeletons require endless manual adjustments — turning knobs, calibrating settings, and tweaking controls. 

“It can be frustrating, even nearly impossible, to get it right for each person,” said Aaron Young, associate professor in the George W. Woodruff School of Mechanical Engineering. “With AI, the exoskeleton figures out the mapping itself. It learns the timing of someone’s gait through a neural network, without an engineer needing to hand-tune everything.”

The software monitors each step, instantly updates, and fine-tunes the support it provides. Over time, the exoskeleton aligns its movements with the unique gait of the person wearing it. In this study, the research team used a hip exoskeleton, which provides torque at the hip joint — in other words, adding power to help stroke survivors walk or move their legs more easily.
 

Taking Smarter Steps

Walking after a stroke can be tough and unpredictable. A patient’s stride can change from one day to the next, and even from one step to the next. Most exoskeletons aren’t built for that kind of variation. They are designed around the steady, even gait of healthy young adults, which can leave stroke survivors feeling more unsteady than supported.

Young’s breakthrough, detailed in IEEE Transactions on Robotics, is a neural network — a type of AI that learns patterns much like the human brain does. Sensors at the hip pick up how someone is moving, and the network translates those signals into just the right boost of power to support each step. It quickly figures out a person’s unique walking pattern. But lead clinician Kinsey Herrin said the AI’s learning doesn’t stop there. It keeps adjusting as the patient walks, so the exoskeleton can stay in sync even during stride shifts.

“The speed really surprised us,” Young said. “In just one to two minutes of walking, the system had already learned a person’s gait pattern with high accuracy. That’s a big deal, to adapt that quickly and then keep adapting as they move.”

Tests showed the system was far more accurate than the standard exoskeleton. It reduced errors in tracking stroke patients’ walking patterns by 70%.

Young emphasized that this research is about more than metrics. “When you see someone able to walk farther without becoming exhausted, that’s when you realize this isn’t just about robotics — it’s about giving people back a measure of independence,” he said.
 

Adapting Anywhere

Every exoskeleton comes with its own set of sensors, so the data they collect can look completely different from one device to the next. A neural network trained on one machine often stumbles when it’s moved to another. To get around that, Young’s team designed software that works like a universal adapter plug — no matter what device it’s connected to, it converts the signals into a form the AI can use. After just 10 strides of calibration, the system cut error rates by more than 75%.

“The goal is that someone could strap on a device, and, within a minute, it feels like it was built just for them,” Young said.


A Step Toward the Future

While the study centered on stroke survivors, the implications are far broader. The same adaptive approach could support older adults coping with age-related muscle weakness, people with conditions like Parkinson’s or osteoarthritis, or even children with neurological disabilities. 
Young and his team are now running clinical trials to measure how well the AI-powered exoskeleton supports people in a wide range of everyday activities.

“There’s no such thing as an ‘average’ user,” Young said. “The real challenge is designing technology that can adapt to the full spectrum of human mobility.”

If Georgia Tech’s exoskeleton can rise to that challenge, the promise goes well beyond the lab. It could mean a world where technology doesn’t just help people walk — it learns to walk with them.

Inseung Kang, who holds a B.S., M.S., and Ph.D. from Georgia Tech, is the paper’s lead author and now an assistant professor of mechanical engineering at Carnegie Mellon University. He explained that the real promise is in what comes next. 

“We’ve developed a system that can adjust to a person’s walking style in just minutes. But the potential is even greater. Imagine an exoskeleton that keeps learning with you over your lifetime, adjusting as your body and mobility change. Think of it as a robot companion that understands how you walk and gives you the right assistance every step of the way.”

 

Aaron Young is affiliated with Georgia Tech’s Institute for Robotics and Intelligent Machines.

This research was primarily funded by a grant (DP2HD111709-01) from the National Institutes of Health New Innovator Award Program. Georgia Tech researchers have created the first lung-on-a-chip with a functioning immune system, allowing it to respond to infections much like a real human lung. The breakthrough, published in Nature Biomedical Engineering, provides a more accurate way to study diseases, test therapies, and reduce reliance on animal models. With potential applications in conditions from influenza to cancer, the technology opens the door to personalized medicine that predicts how individual patients will respond to treatment.

Danfei Xu, an assistant professor in Georgia Tech’s School of Interactive Computing, is introducing new models that provide robots with “on-the-job” training.

The National Science Foundation (NSF) awarded Xu its CAREER award given to early career faculty. The award will enable Xu to expand his research and refine his models, which could accelerate the process of robot deployment and alleviate manufacturers from the burden of achieving perfection.

“The main problem we’re trying to tackle is how to allow robots to learn on the job,” Xu said. “How should it self-improve based on the performance or the new requirements or new user preferences in each home or working environment? You cannot expect a robot manufacturer to program all of that.

“The challenging thing about robotics is that the robot must get feedback from the physical environment. It must try to solve a problem to understand the limits of its abilities so it can decide how to improve its own performance.”

As with humans, Xu views practice as the most effective way for a robot to improve a skill. His models train the robot to identify the point at which it failed in its task performance.

“It identifies that skill and sets up an environment where it can practice,” he said. “If it needs to improve opening a drawer, it will navigate itself to the drawer and practice opening it.”

The models allow the robot to split tasks into smaller parts and evaluate its own skill level using reward functions. Cooking dinner, for example, can be divided into steps like turning on the stove and opening the fridge, which are necessary to achieve the overall goal.

“Planning is a complex problem because you must predict what’s going to happen in the physical world,” Xu said. “We use machine learning techniques that our group has developed over the past two years, using generated models to generate positive futures. They’re very good at modeling long-horizon phenomena.

“The robot knows when it’s failed because there’s a value that tells it how well it performed the task and whether it received its reward. While we don’t know how to tell the robot why it failed, we have ways for it to improve its skills based on that measurement.” 

One of the biggest barriers that keeps many robots from being made available for public use is the pressure on manufacturers to make the robot as close to perfect as possible at deployment. Xu said it’s more practical to accept that robots will have learning gaps that need to be filled and to implement more efficient real-world learning models.

“We work under the pressure of getting everything correct before deployment,” he said. “We need to meet the basic safety requirements, but in terms of competence, it is difficult to get that perfect at deployment. This takes some of the pressure off because it will be able to self-adapt.”

Virtual Workspace for Data Workers

Yalong Yang, another assistant professor in the School of IC, also received the NSF CAREER Award for a research proposal that will design augmented and virtual reality (AR/VR) workspaces for data workers. 

“In 10 years, I envision everyone will use AR/VR in their office, and it will replace their laptop or their monitor,” Yang said.

Yang said he is also working with Google on the project and using Google Gemini to bring conventional applications to immersive space, with data tools being the most complicated systems to re-design for immersive environments.

The immersive workspace and interface will also enable teams of data workers to collaborate and share their data in real-time.

“I want to support the end-to-end process,” Yang said. “We have visualization tools for data, but it’s not enough. Data science is a pipeline — from collecting data to processing, visualizing, modeling and then communicating. If you only support one, people will need to switch to other platforms for the other steps.”

Yang also noted that prior research has shown that VR can enhance cognitive abilities, such as memory and attention and support multitasking. The results of his project could lead to maximizing worker efficiency without them feeling strained.

“We all have a cognitive limit in our working memory. Using AR/VR can increase those limits and process more information. We can expand people’s spatial ability to help them build a better mental model of the data presented to them.”

Yang was also recently named a 2025 Google Research Scholar as he seeks to build a new artificial intelligence (AI) tool that converts mobile apps into 3D immersive environments.

Before launching its prototype, a research team within Georgia Tech’s School of Interactive Computing, led by Professor Bruce Walker and Assistant Professor Sehoon Ha, is working to improve its methods and designs based on research within blind and visually impaired (BVI) communities.

“There’s been research on the technical aspects and functionality of robotic guide dogs, but not a lot of emphasis on the aesthetics or form factors,” said Avery Gong, a recent master’s graduate who worked in Walker’s lab. “We wanted to fill this gap.”

Training a guide dog can cost up to $50,000, and while there are nonprofit organizations that can cover these costs for potential owners, there is still a gap between the amount of available guide dogs and BVI individuals who need them. Not all BVI individuals are able to care for a dog and feed it. The dog also has fewer than 10 working years before it needs replacement.

Gong co-authored a paper on the design implications of the robotic guide dog that was presented at the 2025 International Conference on Robotics and Automation (ICRA) in Atlanta in May.

The consensus among the study’s participants indicates they prefer a robotic guide dog that:

• resembles a real dog and appears approachable
• has a clear identifier of being a guide dog, such as a vest
• has built-in GPS and Bluetooth connectivity
• has control options such as voice command
• has soft textures without feeling furry
• has long battery life and self-charging capability

“A lot of people said they didn’t want the dog to look too cute or appealing because it would draw too much attention,” said Aviv Cohav, another lead author of the paper and recent master’s graduate.

“Many people have issues with taking their guide dog to places, whether it’s little kids wanting to play with the dog or people not liking dogs or people being scared of them, and that reflects on the owners themselves. We wanted to look at what would be a good balance between having a functional robot that wouldn’t scare people away or be a distraction.”

The researchers also had to consider the perspectives of sighted individuals and how society at large might view a robotic guide dog.

An example of this is the amount of noise the dog makes while walking. The owner needs to hear the dog is active, but the clanky sound many off-the-shelf robots make could create disturbances in indoor spaces that amplify sounds. To offset the noise, the team developed algorithms that allow the robot to move more quietly.

Walker and his lab have examined similar scenarios that must take public perception into account.

“We like to think of Georgia Tech as going the extra mile,” Walker said. “Let’s not just make a robot, but a robot that’s going to fit into society.

“To have impact, the technologies we produce must be produced with society in mind. This is a holistic design that considers the users and all the people with whom the users interact.”

Taery Kim, a computer science Ph.D. student, began working on the concept of a robotic guide dog when she came to Georgia Tech in 2022. She and Ha, her advisor, have authored papers on building the robot’s navigation and safety components. 

“When I started, I thought it would be as simple as giving the guide dog a command to take me to Starbucks or the grocery store, and it would just take me,” Kim said. “But the user must give waypoint directions — ‘go left here,’ ‘turn right,’ ‘go forward,’ ‘stop.’ Detailed commands must be delivered to the dog.”

While a real dog has naturally enhanced senses of hearing and smell that can’t be replicated, technology can provide interconnected safety features during an emergency. The researchers envision a camera system equipped with a 360-degree field of view, computer vision algorithms that detect obstacles or hazards, and voice recognition that recognizes calls for help. An SOS function could automatically call 911 at the owner’s request or if the owner is unresponsive.

Kim said the robot should also have explainability features to enhance communication with the owner. For example, if the robot suddenly stops or ignores an owner’s commands, it should tell the owner that it’s detecting a hazard in their path.

Manufacturing a robot at scale would initially be expensive, but the researchers believe the cost would eventually be offset because of its longevity. BVI individuals may only need to purchase one during their lifetime.

To introduce a prototype, the multidisciplinary research team recognizes that it needs to enlist experts from other fields to adequately address the various implications and research gaps inherent in the project.

Walker said the teams welcome additional partners who are keen to tackle challenges ranging from design and engineering to battery life to human-robot interaction.

Team member J. Taery Kim was supported by the National Science Foundation's Graduate Research Fellowship Program (NSF GRFP) under Grant No. DGE-2039655.

 Led by PI: Taylor Higgins, Assistant Professor, FAMU-FSU Department of Mechanical Engineering, partnering with Co-PIs Shreyas Kousik, Assistant Professor, Georgia Tech, George W. Woodruff School of Mechanical Engineering, and Brady DeCouto, Assistant Professor, FSU Anne Spencer Daves College of Education, Health, and Human Sciences, the research will use the acquisition of unicycle riding skill by participants to gain a better grasp on human motor learning in tasks requiring balance and complex movement in space. Although it might sound a bit odd, the fact that most people don’t know how to ride a unicycle, and the fact that it requires balance, mean that the data will cover the learning process from novice to skilled across the participant pool.

Using data acquired from human participants, the team will develop a “robotics assistive unicycle” that will be used in the training of the next pool of novice unicycle riders.  This is to gauge if, and how rapidly, human motor learning outcomes improve with the assistive unicycle. The participants that engage with the robotic unicycle will also give valuable insight into developing effective human-robot collaboration strategies.

The fact that deciding to get on a unicycle requires a bit of bravery might not be great for the participants, but it’s great for the research team. The project will also allow exploration into the interconnection between anxiety and human motor learning to discover possible alleviation strategies, thus increasing the likelihood of positive outcomes for future patients and consumers of these devices.

Author
-Christa M. Ernst

This Article Refers to NSF Award # 2449160

Georgia Tech roboticists are attempting to identify the most stressful periods that human teleoperators experience while performing tasks remotely. A novel study provides new insights into determining when a teleoperator needs to operate at a high level of focus and which parts of the task can be delegated to robot automation.

School of Interactive Computing Associate Professor Matthew Gombolay calls it the “sweet spot” of human ingenuity and robotic precision. Gombolay and students from his CORE Robotics Labconducted a novel study that measures stress and workload on human teleoperators.

Gombolay said it can inform military officials on how to strategically implement task automation and maximize human teleoperator performance.

Humans continue to hand over more tasks to robots to perform, but Gombolay said that some functions will still require human input and oversight for the foreseeable future.

Specific applications, such as space exploration, commercial and military aviation, disaster relief, and search and rescue, pose substantial safety concerns. Astronauts stationed on the International Space Station, for example, manually control robots that bring in supplies, move cargo, and make structural repairs.

“It’s brutal from a psychological perspective,” Gombolay said.

The question often asked about automating a task in these fields is, at what point can a robot be trusted more than a human?

A recent paper by Gombolay and his current and former students — Sam Yi Ting, Erin Hedlund-Botti, and Manisha Natarajan — sheds new light on the debate. The paper was published in the IEEE Robotics and Automation Letters and will be presented at the International Conference on Robotics and Automation in Atlanta.

The NASA-funded study can identify which aspects of tedious, time-consuming tasks can be automated and which require human supervision. If roboticists can pinpoint the elements of a task that cause the least stress, they can automate these components and enable humans to oversee the more challenging aspects.

“If we’re talking about repetitive tasks, robots do better with that, so if you can automate it, you should,” said Ting, a former grad student and lead author of the paper. “I don’t think humans enjoy doing repetitive tasks. We can move toward a better future with automation.”

Military officials, for example, could measure the stress of remote drone pilots and know which times during a pilot’s shift require the highest level of attention.

“We can get a sense of how stressed you are and create models of how divided your attention is and the performance rate of the tasks you’re doing,” Gombolay said.

“It can be a low-stress or high-stress situation depending on the stakes and what’s going on with you personally. Are you well-caffeinated? Well-rested? Is there stress from home you’re bringing with you to the workplace? The goal is to predict how good your task performance will be. If it indicates it might be poor, we may need to outsource work to other people or create a safe space for the operator to destress.”

The Stress Test

For their study, the researchers cut a small river-shaped path into a medium-density fiberboard. The exercise required the 24 participants to use a remote robotic arm to navigate through the path from one end to the other without touching the edges.

The experiment grew more challenging as new stress conditions and workload requirements were introduced. The changing conditions required the test participants to multitask to complete the assignment.

Gombolay said the study supports the Yerkes-Dodson Law, which states that moderate levels of stress increase human performance.

The experiment showed that operators felt overwhelmed and performed poorly when multitasking was introduced. Too much stress led to poor performance, but a moderate amount of stress induced more engagement and enhanced teleoperator focus. 

Ting said finding that ideal stress zone can lead to a higher performance rating. 

“You would think the more stressed you are, the more your performance decreases,” Ting said. “Most people didn’t react that way. As stress increased, performance increased, but when you increased workload and gave them more to do, that’s when you started seeing deteriorating performance.”

Gombolay said no stress can be just as detrimental as too much stress. Performing a task without stress tends to cause teleoperators to become disinterested, especially if it is repetitive and time-consuming.

“No stress led to complacency,” Gombolay said. “They weren’t as engaged in completing the task.

“If your excitement is too low, you get so bored you can’t muster the cognitive energy to reason about robot operation problems.”

The Human Factor

Roboticists have made significant leaps in recent years to remove teleoperators from the equation. Still, Gombolay said it’s too early to tell whether robots can be trusted with any task that a human can perform.

“We’re a long way from full autonomy,” he said. “There’s a lot that robots still can’t do without a human operator. Search and rescue operations, if a building collapses, we don’t have much training data for robots to go through rubble by themselves to rescue people. There are ethical needs for humans to be able to supervise or take direct control of robots.”

Top South Korean research institutes have enlisted Georgia Tech researchers Sehoon Ha and Jennifer G. Kim to develop artificial intelligence (AI) to help the humanoid assistant navigate hospitals and interact with doctors, nurses, and patients.

Ha and Kim will partner with Neuromeka, a South Korean robotics company, on a five-year, 10 billion won (about $7.2 million US) grant from the South Korean government. Georgia Tech will receive about $1.8 million of the grant.

Ha and Kim, assistant professors in the School of Interactive Computing, will lead Tech’s efforts and also work with researchers from the Korea Advanced Institute of Science and Technology and the Electronics and Telecommunications Research Institute.

Neuromeka has built industrial robots since its founding in 2013 and recently decided to expand into humanoid service robots.

Lee, the group leader of the humanoid medical assistant project, said he fielded partnership requests from many academic researchers. Ha and Kim stood out as an ideal match because of their robotics, AI, and human-computer interaction expertise. 

For Ha, the project is an opportunity to test navigation and control algorithms he’s developed through research that earned him the National Science Foundation CAREER Award. Ha combines computer simulation and real-world training data to make robots more deployable in high-stress, chaotic environments. 

“Dr. Ha has everything we want to put into our system, including his navigation policies,” Lee said. “He works with robots and AI, and there weren’t many candidates in that space. We needed a collaborator who can create the software and has experience running it on robots.”

Ha said he is already considering how his algorithms could scale beyond hospitals and become a universal means of robot navigation in unstructured real-world environments.

“For now, we’re focusing on a customized navigation model for Korean environments, but there are ways to transfer the data set to different environments, such as the U.S. or European healthcare systems,” Ha said. 

“The final product can be deployed to other systems and industries. It can help industrial workers at factories, retail stores, any place where workers can get overwhelmed by a high volume of tasks.”

Kim will focus on making the robot’s design and interaction features more human. She’ll develop a large-language model (LLM) AI system to communicate with patients, nurses, and doctors. She’ll also develop an app that will allow users to input their commands and queries. 

“This project is not just about controlling robots, which is why Dr. Kim’s expertise in human-computer interaction design through natural language was essential.,” Lee said. 

Kim is interviewing stakeholders from three South Korean hospitals to identify service and care pain points. The issues she’s identified so far relate to doctor-patient communication, a lack of emotional support for patients, and an excessive number of small tasks that consume nurses’ time.

“Our goal is to develop this robot in a very human-centered way,” she said. “One way is to give patients a way to communicate about the quality of their care and how the robot can support their emotional well-being.

“We found that patients often hesitate to ask busy nurses for small things like getting a cup of water. We believe this is an area a robot can support.”

The robot’s hardware will be built in Korea, while Ha and Kim will develop the software in the U.S.

Jong-hoon Park, CEO of Neuromeka, said in a press release the goal is to have a commercialized product as soon as possible. 

“Through this project, we will solve problems that existing collaborative robots could not,” Park said. “We expect the medical AI humanoid robot technology being developed will contribute to reducing the daily work burden of medical and healthcare workers in the field.”

Meta announced in February its partnership with the labs of professors Danfei Xu and Judy Hoffman on a novel computer vision-based algorithm called EgoMimic. It enables robots to learn new skills by imitating human tasks from first-person video footage captured by Meta’s Aria smart glasses. 

Xu’s Robot Learning and Reasoning Lab (RL2) displayed EgoMimic in action at ICRA May 19-23 at the World Congress Center in Atlanta.

Lawrence Zhu, Pranav Kuppili, and Patcharapong “Elmo” Aphiwetsa — students from Xu’s lab — used Egomimic to compete in a robot teleoperation contest at ICRA. The team finished second in the event titled What Bimanual Teleoperation and Learning from Demonstration Can Do Today, earning a $10,000 cash prize.

Teams were challenged to perform tasks by remotely controlling a robot gripper. The robot had to fold a tablecloth, open a vacuum-sealed container, place an object into the container, and then reseal it in succession without any errors.

Teams completed the tasks as many times as possible in 30 minutes, earning points for each successful attempt.

The competition also offered different challenge levels that increased the points awarded. Teams could directly operate the robot with a full workstation view and receive one point for each task completion. Or, as the RL2 team chose, teams could opt for the second challenge level.

The second level required an operator to control the task with no view of the workstation except for what was provided to through a video feed. The RL2 team completed the task seven times and received double points for the challenge level.

The third challenge level required teams to operate remotely from another location. At this level, teams could earn four times the number of points for each successful task completed. The fourth level challenged teams to deploy an algorithm for task performance and awarded eight points for each completion.

Using two of Meta’s Quest wireless controllers, Zhu controlled the robot under the direction of Aphiwetsa, while Kuppili monitored the coding from his laptop.

“It’s physically difficult to teleoperate for half an hour,” Zhu said. “My hands were shaking from holding the controllers in the air for that long.”

Being in constant communication with Aphiwetsa helped him stay focused throughout the contest.

“I helped him strategize the teleoperation and noticed he could skip some of the steps in the folding,” Aphiwetsa said. “There were many ways to do it, so I just told him what he could fix and how to do it faster.”

Zhu said he and his team had intended to tackle the fourth challenge level with the EgoMimic algorithm. However, due to unexpected time constraints, they decided to switch to the second level the day before the competition due to unexpected time constraints. 

“I think we realized the day before the competition training the robot on our model would take a huge amount of time,” Zhu said. “We decided to go for the teleoperation and started practicing.”

He said the team wants to tackle the highest challenge level and use a training model for next year’s ICRA competition in Vienna, Austria.

ICRA is the world’s largest robotics conference, and Atlanta hosted the event for the third time in its history, drawing a record-breaking attendance of over 7,000.

Wang is a recipient of the 2025 Outstanding Dissertation Award from the Association for Computing Machinery Special Interest Group on Computer-Human Interaction (ACM SIGCHI). The award recognizes Wang for his lifelong work on democratizing human-centered AI.

“Throughout my Ph.D. and industry internships, I observed a gap in existing research: there is a strong need for practical tools for applying human-centered approaches when designing AI systems,” said Wang, now a safety researcher at OpenAI.

“My work not only helps people understand AI and guide its behavior but also provides user-friendly tools that fit into existing workflows.”

[Related: Georgia Tech College of Computing Swarms to Yokohama, Japan, for CHI 2025]

Wang’s dissertation presented techniques in visual explanation and interactive guidance to align AI models with user knowledge and values. The work culminated from years of research, fellowship support, and internships.

Wang’s most influential projects formed the core of his dissertation. These included:

• CNN Explainer: an open-source tool developed for deep-learning beginners. Since its release in July 2020, more than 436,000 global visitors have used the tool.
• DiffusionDB: a first-of-its-kind large-scale dataset that lays a foundation to help people better understand generative AI. This work could lead to new research in detecting deepfakes and designing human-AI interaction tools to help people more easily use these models.
• GAM Changer: an interface that empowers users in healthcare, finance, or other domains to edit ML models to include knowledge and values specific to their domain, which improves reliability.
• GAM Coach: an interactive ML tool that could help people who have been rejected for a loan by automatically letting an applicant know what is needed for them to receive loan approval.
• Farsight: a tool that alerts developers when they write prompts in large language models that could be harmful and misused.  

“I feel extremely honored and lucky to receive this award, and I am deeply grateful to many who have supported me along the way, including Polo, mentors, collaborators, and friends,” said Wang, who was advised by School of Computational Science and Engineering (CSE) Professor Polo Chau.

“This recognition also inspired me to continue striving to design and develop easy-to-use tools that help everyone to easily interact with AI systems.”

Like Wang, Chau advised Georgia Tech alumnus Fred Hohman (Ph.D. CSE 2020). Hohman won the ACM SIGCHI Outstanding Dissertation Award in 2022.

Chau’s group synthesizes machine learning (ML) and visualization techniques into scalable, interactive, and trustworthy tools. These tools increase understanding and interaction with large-scale data and ML models. 

Chau is the associate director of corporate relations for the Machine Learning Center at Georgia Tech. Wang called the School of CSE his home unit while a student in the ML program under Chau.

Wang is one of five recipients of this year’s award to be presented at the 2025 Conference on Human Factors in Computing Systems (CHI 2025). The conference occurs April 25-May 1 in Yokohama, Japan. 

SIGCHI is the world’s largest association of human-computer interaction professionals and practitioners. The group sponsors or co-sponsors 26 conferences, including CHI.

Wang’s outstanding dissertation award is the latest recognition of a career decorated with achievement.

Months after graduating from Georgia Tech, Forbes named Wang to its 30 Under 30 in Science for 2025 for his dissertation. Wang was one of 15 Yellow Jackets included in nine different 30 Under 30 lists and the only Georgia Tech-affiliated individual on the 30 Under 30 in Science list.

While a Georgia Tech student, Wang earned recognition from big names in business and technology. He received the Apple Scholars in AI/ML Ph.D. Fellowship in 2023 and was in the 2022 cohort of the J.P. Morgan AI Ph.D. Fellowships Program.

Along with the CHI award, Wang’s dissertation earned him awards this year at banquets across campus. The Georgia Tech chapter of Sigma Xi presented Wang with the Best Ph.D. Thesis Award. He also received the College of Computing’s Outstanding Dissertation Award.

“Georgia Tech attracts many great minds, and I’m glad that some, like Jay, chose to join our group,” Chau said. “It has been a joy to work alongside them and witness the many wonderful things they have accomplished, and with many more to come in their careers.”

However, this new generation of self-thinking robots will need security protocols to ensure they aren’t susceptible to hackers. To integrate such robots into society, they must come with assurances that they will behave safely around humans.

It begs the question: can you guarantee the safety of something that doesn’t exist yet? It’s something Assistant Professor Glen Chou hopes to accomplish by developing algorithms that will enable autonomous systems to learn and adapt while acting with safety and security assurances.

He plans to launch research initiatives, in collaboration with the School of Cybersecurity and Privacy and the Daniel Guggenheim School of Aerospace Engineering, to secure this new technological frontier as it develops.

“To operate in uncertain real-world environments, robots and other autonomous systems need to leverage and adapt a complex network of perception and control algorithms to turn sensor data into actions,” he said. “To obtain realistic assurances, we must do a joint safety and security analysis on these sensors and algorithms simultaneously, rather than one at a time.”

This end-to-end method would proactively look for flaws in the robot’s systems rather than wait for them to be exploited. This would lead to intrinsically robust robotic systems that can recover from failures.

[RELATED: New Algorithm Teaches Robots Through Human Perspective]

Chou said this research will be helpful in other domains, including advanced space exploration. If a space rover is sent to one of Saturn’s moons, for example, it needs to be able to act and think independently of scientists on Earth. 

Aside from fighting fires and exploring space, this technology could perform maintenance in nuclear reactors, automatically maintain the power grid, and make autonomous surgery safer. It could also bring assistive robots into the home, enabling higher standards of care. 

This is a challenging domain where safety, security, and privacy concerns are paramount due to frequent, close contact with humans.

This will start in the newly established Trustworthy Robotics Lab at Georgia Tech, which Chou directs. He and his Ph.D. students will design principled algorithms that enable general-purpose robots and autonomous systems to operate capably, safely, and securely with humans while remaining resilient to real-world failures and uncertainty.

Chou earned dual bachelor’s degrees in electrical engineering and computer sciences as well as mechanical engineering from the University of California, Berkeley, in 2017, a master’s and Ph.D. in electrical and computer engineering from the University of Michigan in 2019 and 2022, respectively. 

He was a postdoc at the Massachusetts Institute of Technology Computer Science & Artificial Intelligence Laboratory before joining Georgia Tech in November 2024. He received the National Defense Science and Engineering Graduate fellowship program, NSF Graduate Research fellowships, and was named a Robotics: Science and Systems Pioneer in 2022.

Georgia Tech’s Space Exploration and Analysis Laboratory (SEAL) has developed new algorithms that are headed to the Moon, as part of the Intuitive Machine’s IM-2 mission. The mission is sending a Nova-C class lunar lander named Athena to the Moon’s south pole region to test technologies and collect data that aim to enable future exploration. The mission is part of NASA’s Commercial Lunar Payload Services (CLPS) initiative.

Current training methods limit robots from being produced at the necessary scale to put a robot in every home, said Simar Kareer, a Ph.D. student in the School of Interactive Computing.

“Traditionally, collecting data for robotics means creating demonstration data,” Kareer said. “You operate the robot’s joints with a controller to move it and achieve the task you want, and you do this hundreds of times while recording sensor data, then train your models. This is slow and difficult. The only way to break that cycle is to detach the data collection from the robot itself.”

[VIDEO: Meta Shares EgoMimic Case Study Video]

Other fields, such as computer vision and natural language processing (NLP), already leverage training data passively culled from the internet to create powerful generative AI and large-language models (LLMs).

Many roboticists, however, have shifted toward interventions that allow individual users to teach their robots how to perform tasks. Kareer believes a similar source of passive data can be established to enable practical generalized training that scales the production of humanoid robots.

This is why Kareer collaborated with School of IC Assistant Professor Danfei Xu and his Robot Learning and Reasoning Lab to develop EgoMimic, an algorithmic framework that leverages data from egocentric videos.

Meta’s Ego4D dataset inspired Kareer’s project. The benchmark dataset, released in 2023, consists of first-person videos of humans performing daily activities. This open-source data set trains AI models from a first-person human perspective.

“When I looked at Ego4D, I saw a dataset that’s the same as all the large robot datasets we’re trying to collect, except it’s with humans,” Kareer said. “You just wear a pair of glasses, and you go do things. It doesn’t need to come from the robot. It should come from something more scalable and passively generated, which is us.”

Kareer acquired a pair of Meta’s Project Aria research glasses, which contain a rich sensor suite and can record video from a first-person perspective through external RGB and SLAM cameras.

Kareer recorded himself folding a shirt while wearing the glasses and repeated the process. He did the same with other tasks such as placing a toy in a bowl and groceries into a bag. Then, he constructed a humanoid robot with pincers for hands and attached the glasses to the top to mimic a first-person viewpoint.

The robot performed each task repeatedly for two hours. Kareer said building a traditional training algorithm would take days of teleoperating and recording robot sensory data. For his project, he only needed to gather a baseline of sensory data to ensure performance improvement. 

Kareer bridged the gap between the two training sets with the EgoMimic algorithm. The robot’s task performance rating increased by as much as 400% among various tasks with just 90 minutes of recorded footage. It also showed the ability to perform these tasks in unseen environments.

If enough people wear Aria glasses or other smart glasses while performing daily tasks, it can create the passive data bank needed to train robots on a massive scale.

This type of data collection can enable nearly endless possibilities for roboticists to help humans achieve more in their everyday lives. Humanoid robots can be produced and trained at an industrial level and be able to perform tasks the same way humans do.

“This work is most applicable to jobs that you can get a humanoid robot to do,” Kareer said. “In whatever industry we are allowed to collect egocentric data, we can develop humanoid robots.”

Kareer will present his paper on EgoMimic at the 2025 IEEE Engineers’ International Conference on Robotics and Automation (ICRA), which will take place from May 19 to 23 in Atlanta. The paper was co-authored by Xu and School of IC Assistant Professor Judy Hoffman, fellow Tech students Dhruv Patel, Ryan Punamiya, Pranay Mathur, and Shuo Cheng, and Chen Wang, a Ph.D. student at Stanford.

However, this new generation of self-thinking robots would need security protocols to ensure they aren’t susceptible to hackers. To integrate such robots into society, they must come with assurances that they will behave safely around humans.

It begs the question: can you guarantee the safety of something that doesn’t exist yet? It’s something Assistant Professor Glen Chou hopes to accomplish by developing algorithms that will enable autonomous systems to learn and adapt while acting with safety and security assurances. 

He plans to launch research initiatives, in collaboration with the School of Cybersecurity and Privacy and the Daniel Guggenheim School of Aerospace Engineering, to secure this new technological frontier as it develops. 

“To operate in uncertain real-world environments, robots and other autonomous systems need to leverage and adapt a complex network of perception and control algorithms to turn sensor data into actions,” he said. “To obtain realistic assurances, we must do a joint safety and security analysis on these sensors and algorithms simultaneously, rather than one at a time.”

This end-to-end method would proactively look for flaws in the robot’s systems rather than wait for them to be exploited. This would lead to intrinsically robust robotic systems that can recover from failures.

Chou said this research will be useful in other domains, including advanced space exploration. If a space rover is sent to one of Saturn’s moons, for example, it needs to be able to act and think independently of scientists on Earth. 

Aside from fighting fires and exploring space, this technology could perform maintenance in nuclear reactors, automatically maintain the power grid, and make autonomous surgery safer. It could also bring assistive robots into the home, enabling higher standards of care. 

This is a challenging domain where safety, security, and privacy concerns are paramount due to frequent, close contact with humans.

This will start in the newly established Trustworthy Robotics Lab at Georgia Tech, which Chou directs. He and his Ph.D. students will design principled algorithms that enable general-purpose robots and autonomous systems to operate capably, safely, and securely with humans while remaining resilient to real-world failures and uncertainty.

Chou earned dual bachelor’s degrees in electrical engineering and computer sciences as well as mechanical engineering from University of California Berkeley in 2017, a master’s and Ph.D. in electrical and computer engineering from the University of Michigan in 2019 and 2022, respectively. He was a postdoc at MIT Computer Science & Artificial Intelligence Laboratory prior to joining Georgia Tech in November 2024. He is a recipient of the National Defense Science and Engineering Graduate fellowship program, NSF Graduate Research fellowships, and was named a Robotics: Science and Systems Pioneer in 2022.

Most hepatic diseases are chronic conditions that will be present over the life of the patient, but early detection improves overall health and the ability to manage specific conditions over time. Additionally, assessing patients over time allows for effective treatments to be adjusted as necessary. The standard protocol for diagnosis, as well as follow-up tissue assessment, is a biopsy after the return of an abnormal blood test, but biopsies are time-consuming and pose risks for the patient. Several non-invasive imaging techniques have been developed to assess the stiffness of liver tissue, an indication of scarring, including magnetic resonance elastography (MRE).

MRE combines elements of ultrasound and MRI imaging to create a visual map showing gradients of stiffness throughout the liver and is increasingly used to diagnose hepatic issues. MRE exams, however, can fail for many reasons, including patient motion, patient physiology, imaging issues, and mechanical issues such as improper wave generation or propagation in the liver. Determining the success of MRE exams depends on visual inspection of technologists and radiologists. With increasing work demands and workforce shortages, providing an accurate, automated way to classify image quality will create a streamlined approach and reduce the need for repeat scans. 

Professor Jun Ueda in the George W. Woodruff School of Mechanical Engineering and robotics Ph.D. student Heriberto Nieves, working with a team from the Icahn School of Medicine at Mount Sinai, have successfully applied deep learning techniques for accurate, automated quality control image assessment. The research, “Deep Learning-Enabled Automated Quality Control for Liver MR Elastography: Initial Results,” was published in the Journal of Magnetic Resonance Imaging.

Using five deep learning training models, an accuracy of 92% was achieved by the best-performing ensemble on retrospective MRE images of patients with varied liver stiffnesses. The team also achieved a return of the analyzed data within seconds. The rapidity of image quality return allows the technician to focus on adjusting hardware or patient orientation for re-scan in a single session, rather than requiring patients to return for costly and timely re-scans due to low-quality initial images.

This new research is a step toward streamlining the review pipeline for MRE using deep learning techniques, which have remained unexplored compared to other medical imaging modalities.  The research also provides a helpful baseline for future avenues of inquiry, such as assessing the health of the spleen or kidneys. It may also be applied to automation for image quality control for monitoring non-hepatic conditions, such as breast cancer or muscular dystrophy, in which tissue stiffness is an indicator of initial health and disease progression. Ueda, Nieves, and their team hope to test these models on Siemens Healthineers magnetic resonance scanners within the next year.

Publication
Nieves-Vazquez, H.A., Ozkaya, E., Meinhold, W., Geahchan, A., Bane, O., Ueda, J. and Taouli, B. (2024), Deep Learning-Enabled Automated Quality Control for Liver MR Elastography: Initial Results. J Magn Reson Imaging. https://doi.org/10.1002/jmri.29490

Prior Work 
Robotically Precise Diagnostics and Therapeutics for Degenerative Disc Disorder

Related Material
Editorial for “Deep Learning-Enabled Automated Quality Control for Liver MR Elastography: Initial Results”

Since humans often associate objects in a home with the room they are in, Naoki Yokoyama thinks robots that navigate human environments to perform assistive tasks should mimic that reasoning.

Roboticists have employed natural language models to help robots mimic human reasoning over the past few years. However, Yokoyama, a Ph.D. student in robotics, said these models create a “bottleneck” that prevents agents from picking up on visual cues such as room type, size, décor, and lighting. 

Yokoyama presented a new framework for semantic reasoning at the Institute of Electrical and Electronic Engineers (IEEE) International Conference on Robotics and Automation (ICRA) last month in Yokohama, Japan. ICRA is the world’s largest robotics conference.

Yokoyama earned a best paper award in the Cognitive Robotics category with his Vision-Language Frontier Maps (VLFM) proposal.

Assistant Professor Sehoon Ha and Associate Professor Dhruv Batra from the School of Interactive Computing advised Yokoyama on the paper. Yokoyama authored the paper while interning at the Boston Dynamics’ AI Institute.

“I think the cognitive robotic category represents a significant portion of submissions to ICRA nowadays,” said Yokoyama, whose family is from Japan. “I’m grateful that our work is being recognized among the best in this field.”

Instead of natural language models, Yokoyama used a renowned vision-language model called BLIP-2 and tested it on a Boston Dynamics “Spot” robot in home and office environments.

“We rely on models that have been trained on vast amounts of data collected from the web,” Yokoyama said. “That allows us to use models with common sense reasoning and world knowledge. It’s not limited to a typical robot learning environment.”

What is Blip-2?

BLIP-2 matches images to text by assigning a score that evaluates how well the user input text describes the content of an image. The model removes the need for the robot to use object detectors and language models. 

Instead, the robot uses BLIP-2 to extract semantic values from RGB images with a text prompt that includes the target object. 

BLIP-2 then teaches the robot to recognize the room type, distinguishing the living room from the bathroom and the kitchen. The robot learns to associate certain objects with specific rooms where it will likely find them.

From here, the robot creates a value map to determine the most likely locations for a target object, Yokoyama said.

Yokoyama said this is a step forward for intelligent home assistive robots, enabling users to find objects — like missing keys — in their homes without knowing an item’s location. 

“If you’re looking for a pair of scissors, the robot can automatically figure out it should head to the kitchen or the office,” he said. “Even if the scissors are in an unusual place, it uses semantic reasoning to work through each room from most probable location to least likely.”

He added that the benefit of using a VLM instead of an object detector is that the robot will include visual cues in its reasoning.

“You can look at a room in an apartment, and there are so many things an object detector wouldn’t tell you about that room that would be informative,” he said. “You don’t want to limit yourself to a textual description or a list of object classes because you’re missing many semantic visual cues.”

While other VLMs exist, Yokoyama chose BLIP-2 because the model:

• Accepts any text length and isn’t limited to a small set of objects or categories.
• Allows the robot to be pre-trained on vast amounts of data collected from the internet.
• Has proven results that enable accurate image-to-text matching.

Home, Office, and Beyond

Yokoyama also tested the Spot robot to navigate a more challenging office environment. Office spaces tend to be more homogenous and harder to distinguish from one another than rooms in a home. 

“We showed a few cases in which the robot will still work,” Yokoyama said. “We tell it to find a microwave, and it searches for the kitchen. We tell it to find a potted plant, and it moves toward an area with windows because, based on what it knows from BLIP-2, that’s the most likely place to find the plant.”

Yokoyama said as VLM models continue to improve, so will robot navigation. The increase in the number of VLM models has caused robot navigation to steer away from traditional physical simulations.

“It shows how important it is to keep an eye on the work being done in computer vision and natural language processing for getting robots to perform tasks more efficiently,” he said. “The current research direction in robot learning is moving toward more intelligent and higher-level reasoning. These foundation models are going to play a key role in that.”

Top photo by Kevin Beasley/College of Computing.

Georgia Tech’s Simon Sponberg recently collaborated with researchers at the University of Washington, Simon Fraser University, University of Colorado Boulder, and Stanford Research Institute to answer one deceptively complex question: Why can’t robots outrun animals? 

“This work is about trying to understand how, despite have some really amazing robots, there still seems to be a gulf between the capabilities of animal movement and what we can engineer,” says Sponberg, who is Dunn Family Associate Professor in the School of Physics and School of Biological Sciences. 

Recently published in Science Robotics, their study systematically examines a suite of biological and robotic runners to figure out how to further advance our best robotic designs. 

“In robotics design we are often very component focused — we are used to having to establish specifications for the parts that we need and then finding the best component solution,” said Sponberg, who also serves on the executive committee for Georgia Tech's Neuro Next Initiative. “This is of course not how evolution works. We wondered if we systematically analyzed the performance of animals in the same component way that we design robots, if we might see an obvious gap.” 

The gap turns out not to be in the function of individual robotic components, but rather the ability of those components to work together in the seamless way biological components do, highlighting a field of opportunity for new research in robotic development. 

“This means that the frontier is not necessarily figuring out how to design better motors or sensors or controllers,” says Sponberg, “but rather how to integrate them together — this is where biology really excels.” 

Read more about man versus machine and the future of bioinspired robotics here.

Georgia Tech alumna Feifei Qian (M.S. PHYS 2011, Ph.D. ECE 2015), an assistant professor of electrical and computer engineering at the USC Viterbi School of Engineering and School of Advanced Computing, leads the NASA LASSIE project alongside co-investigator Frances Rivera-Hernández, an assistant professor in the School of Earth and Atmospheric Sciences at Georgia Tech. Sharissa Thompson, a graduate student at Georgia Tech, is a student intern on the NASA Curiosity Rover project.

The Palmer Glacier on Oregon’s Mount Hood isn’t the Moon, but it’s a good place to practice.

Some 6,000 feet up the snow-capped mountain, located about 70 miles east of Portland, a multi-disciplinary team from the University of Southern California, Texas A&M University, Georgia Institute of Technology, Oregon State University, Temple University, the University of Pennsylvania, and NASA gathered to turn loose a four-legged robot named Spirit into the wild.

The team that included engineers, cognitive scientists, geoscientists and planetary scientists field-tested Spirit as part of the LASSIE Project: Legged Autonomous Surface Science in Analog Environments. Spirit covered a variety of challenging terrains, using his spindly metal legs to amble over, across and over around shifting dirt, slushy snow and boulders during five days of testing in summer 2023. Sometimes he expertly traversed the hillside, while at other moments he teetered and fell over. All part of the process to better understand the substrate properties and learn to better walk on these extreme terrains. The practice time Spirit logged produced data that will be used to train future robots for use on intergalactic surfaces, like Earth’s moon and perhaps planets in our solar system.

“A legged robot needs to be able to detect what is happening when it interacts with the ground underneath, and rapidly adjust its locomotion strategies accordingly,” says Feifei Qian, an assistant professor of electrical and computer engineering at the USC Viterbi School of Engineering and School of Advanced Computing, which is leading the project funded by NASA. “When the robot leg slips on ice or sinks into soft snow, it inspires us to look for new principles and strategies that can push the boundary of human knowledge and enable new technology. We learn and improve from the observed failures.”

Watch this 5-minute video produced for the team by documentary filmmaker Sean Grasso.

Spirit learns from every step.

“Similar to the way that when we walk on uneven surfaces as humans, we can sort of detect how the ground is shifting beneath our feet, a legged robot is capable of the exact same thing,” says Cristina Wilson, a cognitive scientist at Oregon State University.

The more machines the merrier

Qian’s group doesn’t intend to stop at just one robot, wandering the wilderness alone. She and her former colleagues at Penn, Cynthia Sung, Mark Yim, Daniel Koditschek, and Douglas Jerolmack, received a two-year $2 million grant from NASA they’re calling the TRUSSES Project: Temporarily, Robots Unite to Surmount Sandy Entrapments, Then Separate. They want to help the space agency put teams of robots on the Moon and have them work together on tasks. They would take the knowledge they came in with, and the data they collect on the mission, and communicate those details to each other.

“They would sense how the ground conditions are,” Qian says, “and then exchange that information with one another, and collectively form a map of locomotion risk estimation. The team of robots can then use this traversal risk map to inform their planetary explorations: ‘There is an extremely soft sand patch that might be high-risk for wheeled rovers. Come over here, this might be a safer area.’ ”

The robots in mind for this kind of work would be more than just Spirit: There would be a wheeled rover (great for payload and long distances), a Hexapedal robot (intermediate payload but better mobility than the wheeled), and dog-like ones like the rugged version of Spirit (highest mobility, shorter distances). And here’s the coolest part of that research, the part that sounds like something the Transformers would do. Or at least a team of castaways on “Survivor”: If one got in a jam, made immovable by loose dirt or a rock or a ravine, his bot-mates would arrive and link together and form a bridge, or a pyramid, to hoist their pal to safety. And then back to work.

“When they plan for the strategy to pull the robot up, they’ll decide what force to exert and what position the robot should go to, while also compiling the terrain information,” Qian says. “That’s the key idea of how to use these capabilities: to both prevent and recover from locomotion failures in extreme terrain.”

Back to Mount Hood

Spirit gets around a variety of natural environments, to learn how to better move on challenging terrains. Qian has let him off his leash on Southern California beaches, and the multi-university team has field-tested him in the soft granules of White Sands National Park in New Mexico. But the video shot at Mount Hood shows just how otherworldly that landscape can be in these planetary-analogue environments. This provides Spirit with plenty of opportunities to learn on earth, before potentially exploring other planets.

“You look around us, it would be very hard to drive up this,” Ryan Ewing, a geologist from NASA Johnson Space Center, shares. “But as a legged being, as humans, we can step around it easily. A dog could walk around it easily. So this project is the proving ground that we can enable new science and new mobility on environments that are like other planets.”

In fact, a dog is indeed frisking about: Howard, Wilson’s German shepherd, wandered about, with the kind of agility Spirit could only dream of.

“We are going to observe how Howard moves in different types of snow and ice conditions,” Qian says. “What exactly, out of those combined motions, allows him to succeed on challenging terrain?”

The LASSIE Project calls for two more trips for Spirit: to Mount Hood this summer, and to White Sands next year. The TRUSSES team, from USC and Penn, also plans to visit White Sands next year with Spirit and the other, new, multi-tasking robots. Imagine WALL-E with friends.

—

The NASA PSTAR (Planetary Science and Technology Through Analog Research) number for this project is 80NSSC22K1313.

Seth Hutchinson, executive director of the Institute for Robotics and Intelligent Machines at Georgia Tech, delivered a keynote lecture on “The Impacts of Today’s Robotics Innovation on the Relationship Between Robots and Their Human Co-Workers in Manufacturing Applications” — an overview of current state-of-the-art robotic technologies and future research trends for developing robotics aimed at interactions with human workers in manufacturing.

In addition to the keynote, Hutchinson also participated in the Hyundai Motor Group's Smart Factory Executive Technology Advisory Committee (E-TAC) panel on comprehensive future manufacturing directions and toured the new Hyundai Meta-Factory to observe how digital-twin technology is being applied in their human-robot collaborative manufacturing environment.

Hutchinson is a professor in the School of Interactive Computing. He received his Ph.D. from Purdue University in 1988, and in 1990 joined the University of Illinois Urbana-Champaign, where he was professor of electrical and computer engineering until 2017 and is currently professor emeritus. He has served on the Hyundai Motor Group's Smart Factory E-TAC since 2022.

Hyundai Motor Group Innovation Center Singapore is Hyundai Motor Group’s open innovation hub to support research and development of human-centered smart manufacturing processes using advanced technologies such as artificial intelligence, the Internet of Things, and robotics.

- Christa M. Ernst

Related Links

• Hyundai Newsroom Article: Link
• Event Link: https://mfc2024.com/
• Keynote Speakers: https://mfc2024.com/keynotes/

With a handful of students and researchers at Georgia Tech looking to make breakthroughs in home robotics and object rearrangement, Ramachandruni searched for what others had overlooked.

“To an extent it was challenging, but it was also an opportunity to look at what people are already doing and to get more familiar with the literature,” said Ramachandruni, a Ph.D. student in Robotics. “(Associate) Professor (Sonia) Chernova helped me in deciding how to zone in on the problem and choose a unique perspective.”

Ramachandruni started exploring how a home robot might organize objects according to user preferences in a pantry or refrigerator without prior instructions required by existing frameworks.

His persistence paid off. The 2023 IEEE International Confrence on Robots and Systems (IROS) accepted Ramachandruni’s paper on a novel framework for a context-aware object rearrangement robot.

“Our goal is to build assistive robots that can perform these organizational tasks,” Ramachandruni said. “We want these assistive robots to model the user preferences for a better user experience. We don’t want the robot to come into someone’s home and be unaware of these preferences, rearrange their home in a different way, and cause the users to be distressed. At the same time, we don’t want to burden the user with explaining to the robot exactly how they want the robot to organize their home.”

Ramachandruni’s object rearrangement framework, Context-Aware Semantic Object Rearrangement (ConSOR), uses contextual clues from a pre-arranged environment within its environment to mimic how a person might arrange objects in their kitchen.

“If our ConSOR robot rearranged your fridge, it would first observe where objects are already placed to understand how you prefer to organize your fridge,” he said. “The robot then places new objects in a way that does not disrupt your organizational style.”

The only prior knowledge the robot needs is how to recognize certain objects such as a milk carton or a box of cereal. Ramachandruni said he pretrained the model on language datasets that map out objects hierarchically.

“The semantic knowledge database we use for training is a hierarchy of words similar to what you would see on a website such as Walmart, where objects are organized by shopping category,” he said. “We incorporate this commonsense knowledge about object categories to improve organizational performance.

“Embedding commonsense knowledge also means our robot can rearrange objects it hasn’t been trained on. Maybe it’s never seen a soft drink, but it generally knows what beverages are because it’s trained on another object that belongs to the beverage category.”

Ramachandruni tested ConSOR against two model training baselines. One used a score-based approach that learns how specific users group objects in an environment. It then uses the scores to organize objects for users. The other baseline used the GPT-3 large language model prompted with minimal demonstrations and without fine-tuning to determine the placement of new objects. ConSOR outperformed both baselines.

“GPT-3 was a baseline we were comparing against to see whether this huge body of common-sense knowledge can be used directly without any sort of frame,” Ramachandruni said. “The appeal of LLMs is you don’t need too much data; you just need a small data set to prompt it and give it an idea. We found the LLM did not have the correct inductive bias to correctly reason between different objects to perform this task.”

Ramachandruni said he anticipates there will be scenarios where user input is required. His future work on the project will include minimizing the effort required by the user in those scenarios to tell the robot its preferences.

“There are probably scenarios where it’s just easier to ask the user,” he said. “Let’s say the robot has multiple ideas of how to organize the home, and it’s having trouble deciding between them. Sometimes it’s just easier to ask the user to choose between the options. That would be a human-robot interaction addition to this framework.”

IROS is taking place this week in Detroit.

The CCRL, however, functions as a “library” that allows the robot to remember everything it has learned while performing the simple tasks. Each newly obtained skill is added to the library and leveraged for more complex skills. A turning motion, for instance, can be learned on top of walking while serving as the basis for navigation skills.

Kumar said CCRL has broken new ground on interactive navigation research. Interactive navigation is one of several navigation solutions that allow robots to navigate in the real world. These solutions include point navigation, which trains a robot to reach a point on a map, and object navigation, which teaches it to reach a selected object.

Interactive navigation requires a robot to reach a goal location while interacting with obstacles on the way, which has proven to be the most difficult for robots to learn.

The key, Kumar said, to get a robot to go from walking to pushing an object is in the joints and the robot discovering the different types of motions it can make with them.

So far, Kumar’s policy has reached 10 skills that a robot can learn and deploy. The number of skills it can learn on one policy depends on the hardware the programmer is using.

“It just takes longer to train as you keep adding more skills because now the policy also has to figure out how to incorporate all these skills in different situations,” he said. “But theoretically, you can keep adding more skills indefinitely as long as you have a powerful enough computer to run the policies.”

Kumar said he sees CCRL being useful for home assistant robots, which are required to be agile and limber to navigate around a cluttered household. He also said it could possibly serve as a guide dog for the visually impaired.

“If you have obstacles in front of someone who is visually impaired, the robot can just clear up the obstacles as the person is walking, open the door for them, and things like that,” he said.

NASA plans to launch Lunar Flashlight, a small satellite (SmallSat) about the size of a briefcase that will use lasers to search for water ice inside craters at the Moon’s unexplored South Pole.

NASA says that the Lunar Flashlight, traveling aboard a SpaceX Falcon 9 rocket, will take about three months to reach its “science orbit.” The launch itself has been delayed: SpaceX has pushed back the launch several times. Currently, it is expected to launch later this month. 

The work on earth leading up to the launch has already taken quite some time.

Georgia Tech and GTRI have been instrumental in the development of the Lunar Flashlight mission. Researchers in Georgia Tech’s School of Aerospace Engineering worked with NASA’s Marshall Space Flight Center to develop the SmallSat’s novel propulsion system. Georgia Tech Research Institute (GTRI) collaborated to assemble and test the Lunar Flashlight.

Seasoned researchers were assisted by students in their efforts.

One such student is Mary Kate Broadway, a student assistant in GTRI’s Electro-Optical Systems Laboratory (EOSL), whose academic and professional experiences in modeling and fabrication were called upon to create a near 1:1 model of the Lunar Flashlight SmallSat.

Broadway, who is pursuing a bachelor’s degree in mechatronics, robotics, and automation engineering at Kennesaw State University, used GTRI’s SEEDLab makerspace to fashion the model based on designs produced by NASA.

“I got the SolidWorks (a popular solid modeling computer-aided design and computer-aided engineering application) file, and then I started by taking all the SolidWorks parts, making the 3D printables, and then exporting them out as ‘.stl’ files. Here (at the SEEDLab), I queued everything up and printed it,” Broadway explains. She did “all of the painting and the printing” by herself. "However, of course, the SEEDLab helpers (student assistants) all helped me whenever I had trouble.”

Broadway, who already has a BFA in animation and digital arts from Florida State University, has the savvy to make use of the SEEDLab’s wide variety of equipment.

For the Lunar Flashlight project, Broadway employed:

• An Ultimaker S5 FDM, a fused-filament fabrication 3D printer.
• A FormLabs Cameo resin printer.
• A Glowforge 3D laser printer and cutter.
• Various traditional hand tools.

Broadway employed traditional materials such as PET and PLA plastics for some of the more intricate parts of the model. The main body of the model is aluminum, which Broadway collaborated with the Aero Maker Space on the Georgia Tech campus to get pressed and fashioned to specifications with a Waterjet cutting machine. To simulate working solar panels, Broadway designed printed vinyl labels.

Broadway’s supervisor, EOSL Research Engineer Eric Brown, was initially contacted by Principal Research Engineer Jud Ready, Ph.D., who has worked extensively with NASA. Ready has been the liaison to NASA, reporting on Broadway’s progress.

As of Nov. 4, just days before the Lunar Flashlight launch, Broadway was still engrossed in making final adjustments to the model, particularly the tight tolerances of its solar arrays. Broadway began working on the Lunar Flashlight project in April. Working part-time at the SEEDLab, she has spent dozens of hours—amounting to about a month of work--perfecting the device.

Writer: Christopher Weems

Photos: Sean McNeil

GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

The research, in collaboration with Zoo Atlanta, finds that an elephant’s skin doesn’t uniformly stretch. The top of the trunk is more flexible than the bottom, and the two sections begin to diverge when an elephant reaches more than 10%. When stretching for food or objects, the dorsal section of the trunk slides further forward.  

The findings could improve robotics, which today are typically built for either great strength or flexibility. Unlike an elephant’s trunk, the machines can’t do both.

Read about the study and see video from the experiments. 

The project was devised by robotics Ph.D. student Gerry Chen, in collaboration with Juan-Diego Florez, a fellow graduate student; Frank Dellaert, robotics professor in the School of Interactive Computing, and Seth Hutchinson, professor and KUKA Chair for Robotics. The team’s peer-reviewed study of the robot system was published in the International Conference on Robotics and Automation proceedings for 2022.

How It Works

For a robot to be able to paint in a human style, both the robot and the art must be designed with the other in mind — at least for now. The GTGraffiti system consists of three stages: artwork capture, robot hardware, and planning and control.

First, the team uses motion capture technology to record human artists painting — a strategy that allows for insight into the types of motions required to produce spray-painted artwork. For this study, Chen and the team invited two artists to paint the alphabet in a bubble letter graffiti style. As each artist painted, they recorded the motions of the artist’s hand across the canvas, as well as the movements of the spray paint can itself. Capturing hand and spray paint can trajectories is crucial for the robot to be able to paint using similar layering, composition, and motion as those of a human artist.

The team then processed the data to analyze each motion for speed, acceleration, and size, and used that information for the next stage — designing the robot. Taking these data into consideration, as well as portability and accuracy required for the artwork, they chose to use a cable-driven robot. Cable-driven robots, like the Skycams used in sports stadiums for aerial camera shots, are notable for being able to scale to large sizes. The robot runs on a system of cables, motors, and pulleys. The team’s robot is currently mounted on a 9 by 10-foot-tall steel frame, but Chen says it should be possible to mount it directly onto a flat structure of almost any size, such as the side of a building. 

For the third stage, the artist’s composition is converted into electrical signals. Taken together, the figures form a library of digital characters, which can be programmed in any size, perspective, and combination to produce words for the robot to paint. A human artist chooses shapes from the library and uses them to compose a piece of art. For this study, the team chose to paint the letters “ATL.”

Once the team chooses a sequence and position of characters, they use mathematical equations to generate trajectories for the robot to follow. These algorithmically produced pathways ensure that the robot paints with the correct speed, location, orientation, and perspective. Finally, the pathways are converted into motor commands to be executed.  

With all the computing and competing movements, the motors on the robot could potentially work against each other, threatening to rip the robot apart. To address this, the central robot controller is programmed to recalculate motor commands 1,000 times per second so that the robot can function safely and reliably. Once assembled, the robot can then paint an artwork in the style of a human graffiti artist.

Why Art? Why Graffiti?

Some of the most typical industries for robotics applications include manufacturing, biomedicine, automobiles, agriculture, and the military. But the arts, it turns out, can showcase robotics in an especially powerful way.

“The arts, especially painting or dancing, exemplify some of the most complex and nuanced motions humans can make,” Chen said. “So if we want to create robots that can do the highly technical things that humans do, then creating robots that can dance or paint are great goals to shoot for. These are the types of skills that demonstrate the extraordinary capabilities of robots and can also be applied to a variety of other applications.”  

On a personal level, Chen is motivated by his hope for people to perceive robots as being helpful to humanity, rather than seeing them as job-stealers or entities that cause feelings of fear, sadness, or doom as often depicted in film.

“Graffiti is an art form that is inherently meant to be seen by the masses,” Chen said. “In that respect, I feel hopeful that we can use graffiti to communicate this idea — that robots working together with humans can make positive contributions to society.”

Future Directions

Presently, Chen and the team’s plans for the robot are centered around two main thrusts: preserving and amplifying art. To this end, they are currently experimenting with reproducing pre-recorded shapes at different scales and testing the robot’s ability to paint larger surfaces. These abilities would enable the robot to paint scaled up versions of original works in different geographical locations and for artists physically unable to engage in onsite spray painting. In theory, an artist would be able to paint an artwork in one part of the world, and a GTGraffiti bot could execute that artwork in another place.

In the future, Chen hopes to use GTGraffiti to capture artists painting graffiti in the wild. With the captured motion data, GTGraffiti would be able to reproduce the artwork were it ever painted over or destroyed.

“The robot is not generating the art itself, but rather working together with the human artist to enable them to achieve more than they could without the robot,” Chen said.

Chen envisions that the robot system will eventually have capabilities that allow for real-time artist-robot interaction. He hopes to develop the technology that could enable an artist standing at the foot of a building to spray paint graffiti in a small space while the cable-driven robot copies the painting with giant strokes on the side of the building, for example.

“We hope that our research can help artists compose artwork that, executed by a superhuman robot, communicates messages more powerfully than any piece they could have physically painted themselves,” said Chen.

GTGraffiti is funded by a National Science Foundation grant that supports research involving human-robot collaboration in artistic endeavors.

Citation: G. Chen, S. Baek, J.-D. Florez, W. Qian, S.-W. Leigh, S. Hutchinson, and F. Dellaert, “GTGraffiti: Spray painting graffiti art from human painting motions with a cable driven parallel robot,” in 2022 IEEE International Conference on Robotics and Automation (ICRA), 2022.

DOI: https://doi.org/10.1109/ICRA46639.2022.9812008

Writer: Catherine Barzler

Video: Kevin Beasley

Photography: Rob Felt

Media Contact: Catherine Barzler | catherine.barzler@gatech.edu

ATDC is Georgia’s technology startup incubator and helps entrepreneurs across the state build, launch, and scale successful companies. The goal of the gift is to accelerate growth of automation and robotics by leveraging staff and resources at ATDC in collaboration with Amazon.

“Our mission is to support infrastructure for startups and to help foster compelling startup companies with tremendous talent that solve big problems,” said Thomas Felis, director of robotics strategy for Amazon Global Robotics. “Equally important to us is Georgia Tech’s track record of working with and supporting entrepreneurs from diverse and underrepresented backgrounds.”

The funding includes allocation for an ATDC full-time automation and robotics catalyst to recruit and coach companies focused on automation and robotics. The catalyst will identify relevant startups and help onboard them into ATDC’s startup pipeline and portfolio.

“Georgia Tech is a leader in robotics research, and we are excited to have Amazon support our startup mission at ATDC to bring entrepreneurial ideas to life and to market,” said John Avery, ATDC director. “Innovation can come from anywhere and everywhere, and this collaboration reflects our commitment to support diverse startup founders.”

This effort will also support Georgia Tech’s ongoing robotics research, including the Institute for Robotics and Intelligent Machines.

The Amazon sponsorship expands ATDC’s targeted vertical focus areas to seven, including financial, health, and retail technology, 5G, logistics and supply chain, and advanced manufacturing.

ATDC will also work with Amazon to identify specific areas of technical interest with the aim of developing virtual and physical events to attract relevant startups.

To apply to join the robotics and automation incubator, click here.

This installment of the Faces of Research Q&A series is with Kinsey Herrin

What is your field of expertise and why did you choose it?
I’m a prosthetist/orthotist and conduct research in the field of prosthetics, orthotics/exoskeletons, and rehab robotics. Our goal is to make it easier for people with mobility challenges to live more independent lives by helping them move more easily in the real world. The change we see through our technology sometimes is amazing — people with amputations can go upstairs, step-over-step instead of stiff legged, and kids with walking disabilities start to have more normal walking patterns. As a kid, I always wanted to help people and this profession is the perfect blend of medicine, science, and art — all things that I love plus the added benefit of getting to be around some really incredible people.

What makes Georgia Tech Research institutes unique?
We’re trying to advance technology outside of the lab and into the real world where it can make an impact on real users. That means not only assessing how our users perform with the technology — does it actually make them walk faster, with a more natural and easy gait — but also assessing a user’s own perspective on technology and using all of that data to keep improving the end results. Our facilities and resources are incredible. I often feel I have access to a dream playground for a research prosthetist/orthotist. On top of all of that, our faculty and students are not only extremely talented and at the top of their fields, but I think there is a deeper passion for pursuing this goal to make mobility easier for people with physical challenges.

What impact is your research having on the world? 
I see our work as having an impact on all people with mobility challenges. We are trying to make the world a better place for them by challenging the status quo and saying what clinicians can currently provide is still not good enough. We can still do more to return people to a new normal after amputation, stroke, brain, and spinal cord injuries. When people can access their own environment independently, it has overwhelmingly positive impacts on their quality of life. I think our research is making great strides toward making that possible. 

What do you like to do in your spare time when you are not working on your research or teaching?
I enjoy being outdoors with my husband and son any chance we get. We love pretty much everything about being on or near water — fishing, kayaking, canoeing, swimming, and camping. I also have nine backyard chickens and a dog that are hilarious and fun additions to the Herrin chaos.

“There are swimmer micro-robots that move in a fluid with similar size, but these are the smallest ‘walking’ robots that move on a solid surface,” said Azadeh Ansari, the Sutterfield Family Early Career Assistant Professor at Georgia Tech School of Electrical and Computer Engineering (ECE).

The Georgia Tech study was recently published in the Journal of Micro-Bio Robotics. Currently, most magnetically-actuated micro-bot systems rely on adding multiple electromagnets to enable full control, resulting in higher power consumption and less flexible setups. Being able to demonstrate that a single coil setup is enough for precise bidirectional motion control is a significant hurdle to clear, according to Ansari. With the micro-bots now much easier to operate, the team has been able to demonstrate micromanipulation capabilities.

“With what we’ve shown, we can already think of applying the micro-bots in a lab setting,” said Ansari. “You could have hundreds of robots on the same substrate working akin to ants in a colony.”

In Spring 2019, Ansari’s team showcased larger (two millimeters long) “micro-bristle-bots” that could move by harnessing vibrations. Vibrations are no longer needed to move the micro-bots because of their updated “rocker” design — hence micro-rocker bots. The new design allows the bots to move by performing a stick–slip motion with an out-of-plane magnetic field.

Stick-slip motion basically refers to the two states of the robot; one when the robot is in a pinned/stationary position on the surface and the other when the robot “slips” slightly in one direction and achieves net motion, according to Ph.D. student Tony Wang. When the magnetic field is turned on, the robot will essentially rise and then fall. This motion enables enough kinetic energy to allow the robot to move.

More Than a New Design

Equally as important as the rocker design, the paper demonstrates the novel use of a waveform offset for biasing the direction of the robot's trajectory. The sign of the magnetic field offset (positive or negative), as well as the rocker’s angle with the surface, is what determines the direction the micro-bots will travel. Combined, the rocker design and the magnetic offset make the micro-bots capable of well-controlled, and importantly selectable, movement. The acceleration and deceleration of the micro-rocker bots can further be controlled by changing the frequency of the magnetic field.

The 100-micrometre long micro-bots were 3D printed on to a glass substrate via two-photon lithography and subsequently deposited with a nickel thin film, which acts as a semi-hard magnet in response to external magnetic fields. For many lab applications the robots can be directly printed on the substrate that will go under the microscope, but they can also be printed and transported with a micropipette.

“There are lot of areas the micro-robots can be applied to within the current 2D, under-the-microscope process we’ve established so far,” said Ansari. “But there’s also a future where they can be injected into living organisms to deliver drugs or repair injuries.” 

The team is currently working to equip a micro-bot with a tip that could potentially insert nanoparticles into biological tissue for drug delivery or DNA extraction. Their findings will be presented at the Hilton Head Workshop 2022: A Solid-State Sensors, Actuators and Microsystems Workshop this June.

****

Citation: Tony Wang, DeaGyu Kim, Yifan Shi, and Zhijian Hao, Azadeh Ansari “Bidirectional microscale rocker robots controlled via neutral position offset” (Journal of Micro-Bio Robotics, 2022).  https://doi.org/10.1007/s12213-022-00149-y

Funding: This work is supported by Georgia Tech Institute for Electronics and Nanotechnology (IEN) and the National Science Foundation Graduate Research Fellowship under Grant No. DGE-1650044. The device fabrication was performed at the Georgia Tech Institute for Electronics and Nanotechnology clean room facilities, a member of the National Nanotechnology Coordinated Infrastructure (NNCI), which is supported by the National Science Foundation (Grant ECCS-1542174). 

Active exosuits that incorporate electronics and powered motors are yet to be broadly applied. They tend to be big and heavy, and rely on rigid exoskeletons to transfer weight from body to ground. Exoskeletons add a great deal of stiffness, as well, Young said. Putting on most active exosuits is a little like becoming one with a forklift, restricting a wearer to lifting weights in a vertical plane.

For all these reasons, Young’s Asymmetric Back eXosuit (ABX) described in the October 5 issue of IEEE Transactions on Robotics is highly non-standard. There’s no exoskeleton, no rigid structure, nothing that makes contact with the floor. If the wearer is just standing there, it does nothing except for adding 14 pounds to their legs. But if they raise their body from a leaning over position, it makes a somewhat frantic noise: that is the sound of the ABX helping them rotate their torso, helping them twist. 

Although most active exosuits support vertical lifts, rotating and twisting movements are also ubiquitous, especially in certain fields of manual labor like garbage collection and baggage handling. In many cases, these motions can be awkward and strenuous, leading to work-related injuries as well as back pain, according to Young. Back pain, in turn, is directly correlated with the strength of compressive forces and shear forces that are applied to the spine.

In designing their exosuit, the researchers sought a way to reduce these loads on the spinal joints. Putting it on looks a little like donning a futuristic backpack. Two motors are first strapped onto the back of each upper thigh. These motors are then connected to the back of the opposite shoulders, each with their own cable, making for two cables that diagonally overlap. The exosuit provides assistance by applying tension to the cables when it detects a wearer rise from a bending posture.

“It's definitely a different sensation than a sort of standard exoskeleton. It's not your standard design,” said Young. 

Because the diagonal cables have a component of motion that is horizontal, they exert a pull on the torso that can aid in twisting it from side to side. In tests, the researchers showed that when a wearer of the ABX swung a weight from the ground to one side, the exosuit reduced their back muscle activations by an average of 16%, as measured by electromyography (EMG) sensors. The exosuit also provided a 37% reduction in back muscle exertion when a wearer lifted weights symmetrically, straight off the ground – an assistance level comparable to more rigid designs. 

“People definitely felt like the technology is assisting them, which is great. And we did see the concurrent EMG reduction,” said Young. “I think it’s a great first step.”

In a sense, wearing the exosuit is almost like strapping two additional muscles onto the body – unconventional muscles, which run directly from back to leg. Interestingly, it is the positioning of these muscles rather than their brute strength that makes them functional, said Young.

The motors pull the cables with much less power than the muscles in the body. However, the cables are positioned much further away from the joints. Through this positioning, the cables obtain greater leverage and mechanical advantage, allowing the wearer to reduce their overall muscular output and hence the load that they place on their spine. (Spinal loading was not directly measured in the study.)

Aside from its overall performance, it is the flexible, asymmetric nature of the suit that really makes it unique, Young said. There are currently no other active exosuits that provide assistance for twisting and rotating through a comparable range of motion. While other exosuits also use cables, none have arranged them along diagonal lines.

Young is currently seeking collaborations with industry partners to further develop the exosuit. In future work, he sees its control system as a point to improve. Currently, when a person raises their torso from a lowered position, the cables simply pull with constant tension. But it should be possible to make the system detect different actions of the wearer and adjust its pull in response.

References

J. M. Li, D. D. Molinaro, A. S. King, A. Mazumdar and A. J. Young, "Design and Validation of a Cable-Driven Asymmetric Back Exosuit," in IEEE Transactions on Robotics, doi: 10.1109/TRO.2021.3112280.

About Georgia Tech

The Georgia Institute of Technology, or Georgia Tech, is a top 10 public research university developing leaders who advance technology and improve the human condition. The Institute offers business, computing, design, engineering, liberal arts, and sciences degrees. Its nearly 40,000 students representing 50 states and 149 countries, study at the main campus in Atlanta, at campuses in France and China, and through distance and online learning. As a leading technological university, Georgia Tech is an engine of economic development for Georgia, the Southeast, and the nation, conducting more than $1 billion in research annually for government, industry, and society.

Saad Bhamla, a professor in the School of Chemical and Biomolecular Engineering at Georgia Tech, recently received an Outstanding Investigator Award from the National Institute of General Medical Sciences, part of the National Institutes of Health, to continue studying Spirostomum and attempt to build machines based on similar principles. The grant will provide his research group with $1.98 million in funding over five years. 

For Bhamla, the comparison between the organism and the engine is more than just an analogy. He is now working to build something directly akin to a micro-engine, with pistons and cylinders made out of synthetic cells similar to Spirostomum. 

“That's basically the stuff of my dreams,” Bhamla said. 

Once built, he believes that these molecular engines might prove far more efficient than other miniaturized power sources. The chief difficulty will be making a synthetic cell that functions like Spirostomum, Bhamla said. Today, most synthetic cells do very different things, like producing lab-grown meat. 

“We still think of them as basically bags of fluid,” said Bhamla. “They don't move, they just hang around in test tubes.”

Over the last few years, Bhamla and colleagues have learned more about how Spirostomum works. Its capabilities come from its use of an unconventional fuel, calcium, rather than adenosine triphosphate (ATP), the molecule that powers most human cells.

In a preprint from 2019, Bhamla and Xinjing Xu, then an undergraduate student at Georgia Tech, figured out exactly what makes the organism contract. They found that when calcium binds to Spirostomum’s skeletal mesh, it forces each cell of the skeleton to coil tight.

One of Bhamla’s current doctoral students, Xiangting Lei, is already examining how to replicate this mechanism in a synthetic cell. She is investigating how to give the cell external triggers so that engineers can make it contract whenever they want. Bhamla plans to use the funds from the grant to hire several more graduate students to study these systems.

The goal is to create modern versions of what were historically known as mechanochemical machines. A rich literature on these chemically-powered machines had been created in the sixties, only to be forgotten, Bhamla said. It seemed to be a classic case of science getting ahead of itself. 

“They didn't have the right optical tools and soft materials to do this,” said Bhamla. “This is a great time to revisit [the research] because I think this time, we might be able to have much more success.” 

About Georgia Tech

The Georgia Institute of Technology, or Georgia Tech, is a top 10 public research university developing leaders who advance technology and improve the human condition. The Institute offers business, computing, design, engineering, liberal arts, and sciences degrees. Its nearly 40,000 students representing 50 states and 149 countries, study at the main campus in Atlanta, at campuses in France and China, and through distance and online learning. As a leading technological university, Georgia Tech is an engine of economic development for Georgia, the Southeast, and the nation, conducting more than $1 billion in research annually for government, industry, and society.

Over 130 million pounds of peaches are produced in Georgia per year, and the Southern staple has a total farm gate value of over $71 million, according to recent estimates.

But cultivating peaches is a complex and manually-intensive process that has put a strain on many farms stretched for time and workers. To solve this problem, the Georgia Tech Research Institute (GTRI) has developed an intelligent robot that is designed to handle the human-based tasks of thinning and pruning peach trees, which could result in significant cost savings for peach farms in Georgia.   

"Most folks are familiar with the harvesting of fruit and picking it up at the market," said Ai-Ping Hu, a GTRI senior research engineer who is leading the robot design project. "But there's actually a lot more stuff that gets done before that point in the cultivation cycle."

By using a LIDAR remote sensing system – which determines distances by targeting an object with a laser and measuring the amount of time it takes for the laser beam to reflect back – and a highly-specialized GPS technology that measures locations as specific as a fraction of an inch, the robot is able to self-navigate through peach orchards while steering clear of obstacles. Once at a peach tree, the robot uses an embedded 3D camera to determine which peaches need to be removed, and removes the peaches using a claw-like device, known as an end effector, that is connected to the end of its arm.   

The robot specifically addresses two key components of the peach cultivation cycle: tree pruning and tree thinning. 

Tree pruning refers to the selective removal of branches prior to the spring growing season, which typically occurs from mid-May to early August, and serves many purposes – including exposing more interior surface areas of the fruit trees to sunlight and removing undesired older growth to enable new growth to better thrive.

Tree thinning, meanwhile, is when small or undeveloped peaches, known as peachlets, are removed from peach trees to allow for bigger and better peaches to grow, Hu explained. 

"If you just let all the peaches grow to maturity, then what you'll end up getting is a tree of really small peaches," Hu said. "What you want to do is have relatively few peaches, but you want the ones that remain to be nice and big and sweet – ones you can actually sell."

But so far, there are no robots on the market that have been able to fully replace humans in the peach cultivation due to peach orchards' unstructured environments, which includes unpredictable weather, uneven terrain, and trees' different shapes and sizes, Hu noted.

"In an orchard, no two trees are ever the same," Hu added. "You could have a sunny day or a really cloudy day – that's going to affect the way the technology on the robot can operate."

"There's no robot in the world right now that can harvest or thin peaches as well as people can," Hu said. "The technology's not quite there yet."

Current efforts to automate the harvesting of peaches and other specialty crops so far have not been as successful as advancements in commodity crop automation, where machines can collect hundreds of acres of the good at a time. Commodity crops include items such as corn, wheat, and soybeans. 

"Specialty crops are still very reliant on manual labor," Hu said. "That's really different from something like wheat, where one person driving a combine can harvest thousands of acres, hundreds of acres. Whereas for [peach harvesting], because everything is so individualized so unique, it's really been difficult to automate."

To address these unique issues, GTRI is exploring ways to incorporate artificial intelligence and deep learning training methods to improve the robot's image classification abilities and overall performance. GTRI has also partnered with Dario Chavez, an associate professor in the Department of Horticulture at the University of Georgia Griffin Campus in Griffin, Ga., to further explore the intelligent automation of peach farming.

Gary McMurray, a GTRI principal research engineer and division chief of GTRI's Intelligent Sustainable Technologies Division, said the novel robot stands to transform the fruit cultivation process for many farms that have struggled to grow trees that are strong enough to withstand unpredictable environmental conditions. 

"This is something that directly affects the yield of the trees," McMurray said. "It's money in the pocket of the growers."

Writer: Anna Akins

Photo Credit: Ai-Ping Hu

To collect and transport pollen, honey bees mix pollen particles with regurgitated nectar and form it into a pellet, which clings to each of their hind legs. The honey bees then deposit the pellets into a cell within the hive by carefully scraping them off using their other legs. 

The study, from the lab of George W. Woodruff School of Mechanical Engineering Professor David Hu, sought to better understand the mechanics of this process which could inspire new ways to manufacture and manipulate soft materials. Hu holds a joint appointment in the School of Biological Sciences.

The paper, “Biomechanics of Pollen Removal By the Honey Bee,” is published in the Journal of the Royal Society Interface. 

“We measured the viscoelastic material properties of a pollen pellet,” said Marguerite Matherne, a recent Georgia Tech mechanical engineering Ph.D. graduate who now teaches at Northeastern University. “We found that the pellets have a really long relaxation time, which means they remain mostly in a solid form during the transport process. This is good because it keeps the pellet from melting or falling apart from vibration during flight.”

Matherne and the Georgia Tech research team also tried to replicate how honey bees remove the pellets from their hind legs in the lab. They built a device that scraped adhered pollen pellets from bee legs. The invention produced two discoveries. The first was that the honey bees were much more efficient in removing the pellet than the scraping device they built (the device left much more pollen residue on the leg). They also found that slower removal speeds reduce the force and work required to remove pellets under shear stress. 

“If you remove it slowly, you can avoid applying the excessive force required to remove it quickly,” said Hu, Matherne’s former Georgia Tech advisor. “Removing a pollen pellet is like the opposite of ripping off a Band-Aid.”

Matherne said that there are two key components to the efficiency of the honey bees transporting these pellets. First, the pellets are gooey, allowing them to stick to the hind legs. But, she said, the bees also have a special structure on their legs called the corbicula. It’s fringed with long, curved hairs and becomes embedded into the pellet, allowing for adhesion.

In addition, honey bees can collect pollen particles in various shapes and sizes, while also developing a way to transport them. This is different from other species of bees, which only collect and carry specific types of pollen that are similar in size. They also use different transport techniques.

“Honey bees collect from flowers miles and miles away,” said Hu. “The pollen can change in size by a factor of 10. They must collect all these individual particles and bring it back to one place. And they must do a dozen foraging trips each day, all while keeping their bodies clean. They solve it all by this special method they created to exploit the pellet’s soft material properties.”

The research team believes further studies could lead to new developments in medical patches or fastener applications for soft materials.

“It’s kind of like smart gooey Velcro for soft materials,” said Hu. “It could be a fastener and it knows when you’re trying to remove it so that you don’t have to use an excessive amount of force.”

Matherne suggests that it’s also important to understand the pollinating process since 35% of the world’s crop production depends on pollinators.

“Honey bees are really important pollinators,” said Matherne. “If we want to create a world where we can keep up our pollinators, I think it’s important to understand exactly what they’re doing.”

CITATION: Matherne, M., et.al., "Biomechanics of pollen pellet removal by the honey bee." (Journal of the Royal Society Interface) https://doi.org/10.1098/rsif.2021.0549

As Melkote assumes his appointment, Georgia Tech commends George W. Woodruff School of Mechanical Engineering Regents Professor Surya Kalidindi’s service as the inaugural interim executive director during the Novelis Innovation Hub’s first two years.

Since its establishment in 2019, the Novelis Innovation Hub has set a bold vision to foster world-class partnerships and collaborated with the Institute on battery research, electronics, robotics, high-throughput research, and additive manufacturing.

Advancing Mobility and Sustainability Goals

With additional investment and a permanent leadership appointment to guide the Innovation Hub, Novelis hopes to further advance its position in the aluminum industry through innovation in new technology and application domains, including sustainable mobility, electronics, advanced manufacturing, and supply chain.

“Sustainability is an important element of what Novelis wants to accomplish,” said Melkote, noting Novelis’s target to reduce its carbon footprint by 30% by 2026 and to be net carbon neutral by 2050. “Georgia Tech is focused on a lot of basic science, technologies, and business practices relevant to enabling a more sustainable enterprise.”

Melkote is uniquely qualified for the role, having led the Georgia Tech-Boeing Strategic University Partnership for the last eight years while serving as associate director of Georgia Tech Manufacturing Institute (GTMI). He facilitated the establishment of the Boeing Manufacturing Development Center, an on-campus lab where students and faculty regularly collaborate with a resident Boeing engineer.

“I see this as an opportunity to leverage my experience and knowledge from the Boeing partnership and to expand it. Novelis is engaged in the entire lifecycle of innovation, from early-stage basic research, to applied research and commercialization that will impact society at large,” said Melkote, who also holds the Morris M. Bryan, Jr. Professorship in Mechanical Engineering at Georgia Tech. He will work closely with Dr. Raj Gopalaswamy, Novelis’ global technology director for new domains, who will lead Novelis’ engagement with Georgia Tech.

“To keep advancing the aluminum industry toward the circular economy, we must increase the pace of innovation and develop new solutions that demonstrate aluminum’s superior sustainability benefits,” said Gopalaswamy.  “Through research partnerships with world-leading institutions like Georgia Tech, we can fulfill the growing needs for aluminum applications that help our customers meet their sustainability goals faster and more efficiently.”

Melkote agreed, adding, “What’s exciting is that  Novelis wants to look at the cutting edge of research and see how they can leverage that knowledge to innovate and develop new products.”

“We’re thrilled to have Professor Melkote take on this leadership position in our growing collaboration with Novelis,” said Julia Kubanek, vice president for Interdisciplinary Research at Georgia Tech. “He brings substantial experience to this new role, having built Georgia Tech’s partnership with Boeing and served as associate director of the Georgia Tech Manufacturing Institute for several years.”

Kubanek added that Melkote is well positioned to help Novelis broaden its relationship with Georgia Tech faculty and students, while engaging in key research areas to accelerate Novelis’s product innovation. Additionally, the Innovation Hub intends to not only fund research, but also establish a Scholars Program to fund research fellowships for Georgia Tech graduate and undergraduate students.

“Novelis’s philanthropy commitment allows us to innovate on the educational front, where we can make investments that benefit both Georgia Tech and our educational mission,” said Melkote. “In doing so, we help train the next generation of engineers who will go on to work for companies like Novelis that are committed to sustainability.”

***

About Georgia Tech

The Georgia Institute of Technology, or Georgia Tech, is a top 10 public research university developing leaders who advance technology and improve the human condition. The Institute offers business, computing, design, engineering, liberal arts, and sciences degrees. Its nearly 40,000 students representing 50 states and 149 countries, study at the main campus in Atlanta, at campuses in France and China, and through distance and online learning. As a leading technological university, Georgia Tech is an engine of economic development for Georgia, the Southeast, and the nation, conducting more than $1 billion in research annually for government, industry, and society.

Those so-called terrestrial analog sites on Earth helped NASA choose Jezero for the mission. “Ancient lake beds are a major target for Mars exploration, as they provide evidence for sustained liquid water in Mars’ past — and lake muds commonly preserve biosignatures on Earth,” says Frances Rivera-Hernández, assistant professor in the School of Earth and Atmospheric Sciences. “Thus, if life ever persisted on early Mars, their past presence may be preserved in ancient lake beds.”

Rivera-Hernández, who joined Georgia Tech in January, will soon get a chance to study another analog site in the Antarctic, thanks to a four-year $700,000 NASA grant awarded to her research proposal, “Paleolake deposits in Miers Valley, Antarctica: An analog depositional record for Martian lakes through late Noachian to early Hesperian climatic transitions.”

Just like the drilling and sampling now going on at Jezero Crater on Mars, Rivera-Hernández’s work may help NASA choose future Mars destinations for both robotic rover and crewed missions. That’s because Rivera-Hernández is also a collaborating scientist on NASA’s Curiosity Rover mission. “Lessons learned through the Antarctic project will help inform my work on the mission, as we have been characterizing lake bed deposits with the Rover,” she says. Since landing on Mars in 2012, Curiosity has traveled nearly 26 km (16.14 miles) around the rim of Gale Crater, another probable dry lake.

“I was ecstatic to hear that my grant was funded, and excited to be heading to Antarctica for field work,” says Rivera-Hernández, who will serve as the study’s principal investigator. Her co-investigator is Tyler Mackey, an assistant professor at the University of New Mexico. The grant will also provide funding for two graduate students, one from each institution. Field work is planned to start in January 2024.

“Before the field season, we will be performing remote sensing observations of our field site and performing lab-based analyses on modern lake samples to plan for the field work studying ancient lake beds,” Rivera-Hernández says. 

Her lab team will study the deposits of a large Antarctic lake that persisted through climate changes 10,000 to 20,000 years ago to better recognize those similar changes in ancient lake beds on Mars, like those being explored by Curiosity and Perseverance. 

“Currently, liquid water is not stable on the surface of Mars, but we have abundant geologic evidence for the presence of lakes on early Mars, suggesting that Mars’ climate was different in the past and that it changed through time,” Rivera-Hernández says. “But we still do not have a good understanding on whether this climatic transition was abrupt or gradual, or if Mars was significantly warmer when the lakes were present.”

That’s an unknown because lakes can form in a variety of climates, she adds. Examples are found in polar regions on Earth, where liquid water exists in lakes with permanent ice covers. “However, when ice is present in a lake, there are processes that are unique, and sometimes these produce deposits that may be recorded in lake beds. Thus, past climate may be inferred from lake beds if these unique deposits are recognized and distinguished from other deposit types.”

Rivera-Hernadez’s project will also help scientists recognize these unique deposits in ancient lake beds on Mars — by studying the deposits of that ancient Antarctic lake which experienced periods with and without an ice cover, due to those climatic changes on Earth.

Today, Georgia Tech received two National Science Foundation (NSF) Artificial Intelligence Research Institutes awards, totaling $40 million. A third award for $20 million was granted to the Georgia Research Alliance (GRA), with Georgia Tech serving as one of the leading academic institutions.

“It is essential that we bring together our best minds to ensure that AI delivers on its promise to create a more prosperous, sustainable, safe, and fair future for everyone,” said Ángel Cabrera, president of Georgia Tech. “These NSF awards recognize Georgia Tech’s vast expertise in machine learning and AI and will help us further develop our resources and amplify our impact in these crucial fields.”

Chaouki T. Abdallah, executive vice president for Research at Georgia Tech, concurred, citing major efforts under development to help create a more robust and inclusive future of AI, both on campus and beyond.

“We are incredibly grateful to the NSF for their investment and excited for the opportunities made possible because of this research,” he said. “At Tech, our mission is to advance technology and improve the human condition, catalyzing research that matters. We invested in a unified approach to interdisciplinary research aligned with industry relevance and societal impact, and these awards demonstrate a clear return on that strategy.”

Collectively, NSF made a $220 million investment in 11 new NSF-led Artificial Intelligence Research Institutes.

“I am delighted to announce the establishment of new NSF National AI Research Institutes as we look to expand into all 50 states,” said National Science Foundation Director Sethuraman Panchanathan. “These Institutes are hubs for academia, industry, and government to accelerate discovery and innovation in AI. Inspiring talent and ideas everywhere in this important area will lead to new capabilities that improve our lives, from medicine to entertainment to transportation and cybersecurity, and position us in the vanguard of competitiveness and prosperity.”

Led by NSF, and in partnership with the U.S. Department of Agriculture’s National Institute of Food and Agriculture, the U.S. Department of Homeland Security, Google, Amazon, Intel, and Accenture, the National AI Research Institutes will act as connections in a broader nationwide network to pursue transformational advances in a range of economic sectors, and science and engineering fields — from food system security to next-generation edge networks. In addition to Georgia Tech and GRA, the University of California San Diego, Duke University, Iowa State University, North Carolina State University, The Ohio State University, and University of Washington are the lead universities included in the 11 AI Institutes.

The AI Institutes at Georgia Tech

The three newly established Institutes will address societal challenges, including home care for aging adults; energy, logistics, and supply chains; sustainability; the widening gap in job opportunities; and changing needs in workforce development.

NSF AI Institute for Collaborative Assistance and Responsive Interaction for Networked Groups (AI-CARING) will seek to create a vibrant discipline focused on personalized, collaborative AI systems that will improve quality of care for the aging. The systems will learn individual models of human behavior and how they change over time and use that knowledge to better collaborate and communicate in caregiving environments. Led by Sonia Chernova, associate professor of interactive computing at Georgia Tech, the AI systems will help a growing population of older adults sustain independence, improve quality of life, and increase effectiveness of care coordination across the care network.

“The AI-CARING Institute builds on our existing strengths in AI and in technology for aging. It will create not only novel solutions, but a new generation of researchers focused on the interaction between the two,” said Charles Isbell, dean and John P. Imlay Jr. Chair in the College of Computing. “Our aim is to build cutting-edge technologies that improve the lives of everyone, and I can’t think of a better example than AI-CARING.”

NSF AI Institute for Advances in Optimization (AI4Opt) will revolutionize decision-making on a large scale – fusing AI and mathematical optimization into intelligent systems that will achieve breakthroughs that neither field can achieve independently. Additionally, it will create pathways from high school to undergraduate and graduate education and workforce development training for AI in engineering that will empower a generation of underrepresented students and teachers to join the AI revolution. Led by Pascal Van Hentenryck, A. Russell Chandler III chair and professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech, AI4Opt will tackle use cases in energy, resilience and sustainability, supply chains, and circuit design and control.

“AI4Opt, with its focus on AI and optimization, will create new pathways for novel tools that allow better engineering applications to benefit society,” said Raheem Beyah, dean of Georgia Tech’s College of Engineering and Southern Company Chair. “This will allow engineers to build higher quality materials, more efficient renewable resources, new computing systems, and more, while also reinforcing the field as a career path for diverse students. The new institute complements the College’s commitment to the integration of AI in engineering disciplines.”

NSF AI Institute for Adult Learning and Online Education (ALOE) will lead the country and the world in the development of novel AI theories and techniques for enhancing the quality of adult online education, making this mode of learning comparable to that of in-person education in STEM disciplines. Together with partners in the technical college systems and educational technology sector, ALOE will advance online learning using virtual assistants to make education more available, affordable, achievable, and ultimately more equitable. This Institute is led by the GRA, with support from Georgia Tech and the University System of Georgia (USG). Ashok Goel, professor in the School of Interactive Computing at Georgia Tech, will serve as executive director.  

“Online education for adults has enormous implications for tomorrow’s workforce,” said Myk Garn, a GRA senior advisor, assistant vice chancellor for New Models of Learning at the USG, and ALOE’s principal investigator. “Yet, serious questions remain about the quality of online learning and how best to teach adults online. Artificial intelligence offers a powerful technology for dramatically improving the quality of online learning and adult education.”

The Future of AI at Georgia Tech

Georgia Tech is poised to strategically reimagine the future of AI. Currently, 66% of Georgia Tech undergraduate computer science students have an academic concentration in Intelligence, focusing on the top-to-bottom computational models of intelligence. The College of Computing’s recently launched Ph.D. program in machine learning pulls from faculty in all six colleges across the Institute, and many new courses are being developed that teach AI as a tool for science and engineering. Georgia Tech is exploring the potential creation of a school or college of AI within the next five years, further building on its expansive AI and machine learning footprint. The NSF AI Institutes awards will enable all AI-related academic programs to scale and further differentiate Georgia Tech as a leader in AI education.   

Additionally, the awards will expand and complement ongoing AI research efforts at the Georgia Tech Research Institute (GTRI). In the last fiscal year, GTRI received millions of dollars in research awards from the Department of Defense and other sponsors for AI-affiliated research, and currently, many GTRI researchers are focused on AI-affiliated projects.

“As part of Georgia Tech, GTRI will greatly benefit from the advances in AI that will be achieved as a result of these NSF-funded Institutes, helping us further excel in our aim to deliver leading-edge AI research that benefits national security,” said Mark Whorton, GTRI’s chief technology officer. “GTRI is one of the nation’s leading institutes of applied research for national security specifically because of our deep engagement and close affiliation with the academic units of Georgia Tech. AI is a tool we use in conducting larger research objectives, and we believe strongly that these AI Institutes will enable GTRI to put more research into practice.”

“Georgia Tech has for decades now been pursuing new AI technologies, and now leads the way in AI that is responsible to the needs of the humans who use it,” Isbell said. “We have also worked hard to expand access to AI, especially for underrepresented groups. These Institutes will build on that history, expanding both our ability to create new technologies and to train the next generation of innovators. I look forward to watching them grow and develop.”

About the Georgia Institute of Technology

The Georgia Institute of Technology, or Georgia Tech, is a top 10 public research university developing leaders who advance technology and improve the human condition. The Institute offers business, computing, design, engineering, liberal arts, and sciences degrees. Its nearly 40,000 students, representing 50 states and 149 countries, study at the main campus in Atlanta, at campuses in France and China, and through distance and online learning. As a leading technological university, Georgia Tech is an engine of economic development for Georgia, the Southeast, and the nation, conducting more than $1 billion in research annually for government, industry, and society.

About the National Science Foundation

The U.S. National Science Foundation propels the nation forward by advancing fundamental research in all fields of science and engineering. NSF supports research and people by providing facilities, instruments, and funding to support their ingenuity and sustain the U.S. as a global leader in research and innovation. With a fiscal year 2021 budget of $8.5 billion, NSF funds reach all 50 states through grants to nearly 2,000 colleges, universities, and institutions. Each year, NSF receives more than 40,000 competitive proposals and makes about 11,000 new awards. Those awards include support for cooperative research with industry, Arctic and Antarctic research and operations, and U.S. participation in international scientific efforts.

About the Georgia Research Alliance


The Georgia Research Alliance (GRA) helps Georgia’s university scientists do more research and start more companies. By expanding research and entrepreneurship capacity at public and private universities, GRA grows the Georgia economy by driving more investment in the state, developing a high-tech workforce, and strengthening Georgia’s reputation for innovation. For 30 years, GRA has worked in partnership with the University System of Georgia and the Georgia Department of Economic Development to create the companies and jobs of Georgia’s future. Visit GRA.org for more information.

Contact: Georgia Parmelee | georgia.parmelee@gatech.edu | 404.281.7818

We’ve seen robots take to the air, dive beneath the waves and perform all sorts of maneuvers on land. Now, researchers at UC Santa Barbara and Georgia Institute of Technology are exploring a new frontier: the ground beneath our feet. Taking their cues from plants and animals that have evolved to navigate subterranean spaces, they’ve developed a fast, controllable soft robot that can burrow through sand. The technology not only enables new applications for fast, precise and minimally invasive movement underground, but also lays mechanical foundations for new types of robots.

“The biggest challenges with moving through the ground are simply the forces involved,” said Nicholas Naclerio, a graduate student researcher in the lab of UC Santa Barbara mechanical engineering professor Elliot Hawkes and lead author of a paper on the cover of the journal Science Robotics. Whereas air and water offer little resistance to objects moving through them, he explained, the subterranean world is another story.

“If you’re trying to move through the ground, you have to push the soil, sand or other medium out of the way,” Naclerio said.

Fortunately, the natural world provides numerous examples of underground navigation in the form of plants and fungi that build underground networks and animals that have mastered the ability to tunnel directly through granular media. Gaining a mechanical understanding of how plants and animals have mastered subterranean navigation opens up many possibilities for science and technology, according to Daniel Goldman, Dunn Family Professor of Physics at Georgia Tech.

“Discovery of principles by which diverse organisms successfully swim and dig within granular media can lead to development of new kinds of mechanisms and robots that can take advantage of such principles,” he said. “And reciprocally, development of a robot with such capabilities can inspire new animal studies as well as point to new phenomena in the physics of granular substrates.”

The researchers had a good head start with a vine-like soft robot designed in the Hawkes Lab that mimics plants and the way they navigate by growing from their tips, while the rest of the body remains stationary. In the subterranean setting, tip extension, according to the researchers, keeps resisting forces low and localized only to the growing end; if the whole body moved as it grew, friction over the entire surface would increase as more of the robot entered the sand until the robot could no longer move.

Burrowing animals, meanwhile, serve as inspiration for an additional strategy called granular fluidization, which suspends the particles in a fluid-like state and allows the animal to overcome the high level of resistance presented by sand or loose soil. The southern sand octopus, for instance, expels a jet of water into the ground, and uses its arms to pull itself into the temporarily loosened sand. That ability made its way onto the researchers’ robot in the form of a tip-based flow device that shoots air into the region just ahead of the growing end, enabling it to move into that area.

“The biggest challenge we found and what took the longest to solve was when we switched to horizontal burrowing, our robots would always surface,” Naclerio said. Whereas gases or liquids evenly flow over and under a traveling symmetric object, he explained, in fluidized sand, the distribution of forces is not as balanced, and creates a significant lift force for the horizontally travelling robot. “It’s much easier to push the sand up and out of the way than it is to compact it down.”

To understand the robot’s behavior and the largely unexplored  physics of air-aided intrusions, the team took drag and lift measurements as a result of different angles of airflow into from the tip of a solid rod shoved horizontally into sand.

“Frictional force response in granular materials greatly differs from that of Newtonian fluids, as intruding into sand can compact and stress large swaths of terrain in the direction of motion due to high friction,” said Andras Karsai, a graduate student researcher in Goldman’s lab. “To mitigate this, a low-density fluid that lifts and pushes grains away from an intruder will often reduce the net frictional stress it has to overcome.”

Unlike with gas or liquid, where a downward fluid jet would create lift for the travelling object, in sand the downward air flow reduced the lift forces and excavated the sand below the robot’s growing tip. This, combined with inspiration from the sandfish lizard, whose wedge-shaped head favors downward movement, allowed the researchers to modulate the resisting forces and keep the robot moving horizontally without rising out of the sand.

A small, exploratory, soft robot such as this has a variety of applications where shallow burrowing through dry granular media is needed, such as soil sampling, underground installation of utilities and erosion control. Tip extension enables changes in direction, while also allowing the body of the robot to modulate how firmly anchored it is in the medium — control that could become useful for exploration in low gravity environments. In fact, the team is working on a project with NASA to develop burrowing for the moon or even more distant bodies, like Enceladus, a moon of Jupiter.

“We believe burrowing has the potential to open new avenues and enable new capabilities for extraterrestrial robotics,” Hawkes said.

Research for this paper was conducted also by Mason Murray-Cooper, Yasemin Ozkan-Aydin and Enes Aydin at Georgia Institute of Technology.

Funding: This work is supported by the NSF (grant nos. 1637446, 1915445, 1915355, and 1935548), the Army Research Office (grant no. GR10005043), the Packard Foundation, and by an Early Career Faculty grant from NASA’s Space Technology Research Grants Program. The work of Nicholas D. Naclerio is supported by a NASA Space Technology Research Fellowship. Competing interests: Nicholas D. Naclerio and Elliot W. Hawkes are authors of international patent application WO2020060858A1, related to this work. All other authors declare that they have no competing interests. https://doi.org/10.1126/scirobotics.abe2922

During this time, Georgia Institute of Technology professor Jun Ueda and Ph.D. student Waiman Meinhold, along with their collaborators at NITI-ON Co. and Tohoku University in Japan, began to explore how they might contribute. By employing their previously engineered “smart” tendon hammer and developing a mobile app to accompany it, Meinhold, Ueda, and their collaborators devised a system that enables the deep tendon reflex exam to be performed remotely, filling a gap in neurological healthcare delivery.

The deep tendon reflex exam is both a basic and crucial part of neurological assessment and is often the first step in identifying neurological illnesses. The traditional exam consists of two main parts. First, using a silicone hammer, a physician taps on a patient’s tendon to trigger a reflex response. Next, the physician grades the reflex on a numerical scale. To characterize the reflex, a trained physician relies primarily on previous experience, visual cues, and the “feel” of the hammer rebounding in their hand. Until now, the physical act of reflex elicitation has been completely out of reach for telemedicine. Hitting the correct spot on the tendon is crucial and is necessary in order to elicit a proper reflex response.

According to Meinhold and Ueda’s research, a patient’s caretaker or family member may be able to easily step in to assist with this critical component of the neurological exam. They will simply need to obtain the smart tendon hammer and download the accompanying mobile application for data analysis.

To make this advance possible, Meinhold and Ueda modified a standard commercially available reflex hammer by furnishing it with a small wireless Inertial Measurement Unit (IMU) capable of measuring and streaming the hammer’s acceleration data. In the course of their research, Meinhold and Ueda proved that by taking the hammer’s acceleration measurements from on-tendon and off-tendon locations and running them through a classification algorithm, they can reliably distinguish whether or not the hammer has hit the correct spot.

How would this remote exam work, exactly? Equipped with the smart hammer, the lay person uses the app to select which tendon they will test (bicep, Achilles, patellar, etc.), which calls up the pre-programmed “classifier” for that particular tendon. These “classifiers” are basic forms of artificial intelligence that use aggregated acceleration data collected from experiments to categorize each tap into one of two categories: correct or incorrect. The lay person then uses the smart tendon hammer to administer a tap on the patient’s tendon. As contact is made, the hammer streams acceleration data via Bluetooth to the app, which interprets the data and gives instant feedback to the user about whether they have tapped the correct location. In addition, colored LEDs on the hammer indicate a tap’s success, with a green light indicating a correct tap and a red light indicating an incorrect tap. The user is prompted to keep tapping until they log several correct taps.

Crucially, Meinhold and Ueda showed that lay people can adequately perform tendon tapping. Their research appeared in the peer-reviewed journal Frontiers in Robotics and AI on March 16, 2021. There, moving their smart hammer closer to clinical implementation, Meinhold and Ueda directly compared the manual tapping variability between a novice and a trained clinician. The results were reassuring. The team found that while novices had more variability in their tapping than clinicians, their skill level was adequate. They reliably elicited tendon reflexes. Their research demonstrates that a tool is within reach to allow for remote implementation of deep tendon reflex exam.

But could lay users also aid in grading reflexes? The work by Meinhold and Ueda suggests that non-experts may be able to help. To investigate this, they tested a simple training scheme. They provided participants and physicians with a training video on how to grade reflexes, and then assigned unlabeled videos for them to score. They found that while novices were able to grade reflexes with relatively low error rates, expert physicians outperformed them. Physicians excelled at grading from video, making no errors. To access this expert grading, Meinhold and Ueda envision that through the app, lay users could upload videos of the tendon tapping and reflex response. Physicians could then easily grade the patient’s reflexes from their office.

By revolutionizing a traditional neurological assessment procedure, the smart hammer system developed at Georgia Tech is poised to kick-start a new wave in telemedicine.

Text - Catherine Barzler
Images – Christa Ernst

A Smart Tendon Hammer System for Remote Neurological Examination
W. Meinhold, Y.Yamakawa, H. Honda, T. Mori, S. Izumi and Jun Ueda
Fontiers in Robotics and AI, #8, 2021
DOI=10.3389/frobt.2021.618656       

A research team at the Georgia Institute of Technology has developed a modular solution for handling larger packages without the need for a complex fleet of drones of varying sizes. By allowing teams of small drones to collaboratively lift objects using an adaptive control algorithm, the strategy could allow a wide range of packages to be delivered using a combination of several standard-sized vehicles.

Beyond simplifying the drone fleet, the work could provide more robust drone operations and reduce the noise and safety concerns involved in operating large autonomous UAVs in populated areas. In addition to commercial package delivery, the system might also be used by the military to resupply small groups of soldiers in the field.

“A delivery truck could carry a dozen drones in the back, and depending on how heavy a particular package is, it might use as many as six drones to carry the package,” said Jonathan Rogers, the Lockheed Martin Associate Professor of Avionics Integration in Georgia Tech’s Daniel Guggenheim School of Aerospace Engineering. “That would allow flexibility in the weight of the packages that could be delivered and eliminate the need to build and maintain several different sizes of delivery drones.”

The research was supported, in part, by a National Science Foundation graduate student fellowship and by the Hives independent research and development program of the Georgia Tech Research Institute. A paper on the research has been submitted to the Journal of Aircraft.

A centralized computer system developed by graduate student Kevin Webb would monitor each of the drones lifting a package, sharing information about their location and the thrust being provided by their motors. The control system would coordinate the issuance of commands for navigation and delivery of the package.

“The idea is to make multi-UAV cooperative flight easy from the user perspective,” Rogers said. “We take care of the difficult issues using the onboard intelligence, rather than expecting a human to precisely measure the package weight, center of gravity, and drone relative positions. We want to make this easy enough so that a package delivery driver could operate the system consistently.”

The challenges of controlling a group of robots connected together to lift a package is more complex in many ways than controlling a swarm of robots that fly independently.

“Most swarm work involves vehicles that are not connected, but flying in formations,” Rogers said. “In that case, the individual dynamics of a specific vehicle are not constrained by what the other vehicles are doing. For us, the challenge is that the vehicles are being pulled in different directions by what the other vehicles connected to the package are doing.” 

The team of drones would autonomously connect to a docking structure attached to a package, using an infrared guidance system that eliminates the need for humans to attach the vehicles. That could come in handy for drones sent to retrieve packages that a customer is returning. By knowing how much thrust they are producing and the altitude they are maintaining, the drone teams could even estimate the weight of the package they’re picking up.

Webb and Rogers have built a demonstration in which four small quadrotor drones work together to lift a box that’s 2 feet by 2 feet by 2 feet and weighs 12 pounds. The control algorithm isn’t limited to four vehicles and could manage “as many vehicles as you could put around the package,” Rogers said.

For the military, the modular cargo system could allow squads of soldiers at remote locations to be resupplied without the cost or risk of operating a large autonomous helicopter. A military UAV package retrieval team could be made up of individual vehicles carried by each soldier.

“That would distribute a big lifting capability in smaller packages, which equates to small drones that could be used to team up,” Rogers said. “Putting small drones together would allow them to do bigger things than they could do individually.”

Bringing multiple vehicles together creates a more difficult control challenge, but Rogers argues the benefits are worth the complexity. “The idea of having multiple machines working together provides better scalability than building a larger device every time you have a larger task,” he said. “We think this is the right way to fill that gap.”

Using multiple drones to carry a heavy package could also allow more redundancy in the delivery system. Should one of the drones fail, the others should be able to pick up the load – an issue managed by the central control system. That part of the control strategy hasn’t yet been tested, but it is part of Rogers’ plan for future development of the system.

More research is also needed on the docking system that connects the drones to packages. The structures will have to be made strong and rigid enough to connect to and lift the packages, while being inexpensive enough to be disposable.

“I think the major technologies are already here, and given an adequate investment, a system could be fielded within five years to deliver packages with multiple drones,” Rogers said. “It’s not a technical challenge as much as it is a regulatory issue and a question of societal acceptance.”

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contacts: John Toon (404-894-6986) (jtoon@gatech.edu) or Anne Wainscott-Sargent (404-435-5784) (asargent7@gatech.edu).

Writer: John Toon

Founded in 2015 and now in its second wave of investment with top universities, TRI will invest more than $75 million over the next five years. The university partners will focus on breakthroughs around tough technological challenges in key research priority areas of automated driving, robotics, and machine-assisted cognition.

“Georgia Tech is honored to work closely with TRI to advance robotics in key fields. It’s an exciting start to what we hope will be a longer-term collaboration,” said Seth Hutchinson, executive director of Georgia Tech’s Institute for Robotics and Intelligent Machines and professor and KUKA Chair for Robotics in the School of Interactive Computing. 

"This new phase of university research is about pushing even further and doing so with a broader, more diverse set of stakeholders. To get to the best ideas, collaboration is critical. And we sought out universities like Georgia Tech that share our vision of using AI for human amplification and societal good. The funded projects will contribute to two TRI focus areas: automated driving and home robotics," said Eric Krotkov, TRI chief science officer.

The two Georgia Tech projects total $2.2M over the next three years. Under the agreement, each team will be paired with TRI researchers, who will serve as co-investigators. 

An Outdoor MiniCity to Test Autonomous Driving    
The first research project aims to make it easier for universities to test autonomous vehicles, building on Georgia Tech’s AutoRally platform. Georgia Tech researchers use this small-scale autonomous dirt track to test aggressive driving. The car can control turns and calculate for on-course obstacles at speeds approaching 20 miles per hour. The software and simulation environment could help make future self-driving cars safer under similar hazardous road conditions. Georgia Tech researchers will build on this platform to develop a scale-model MiniCity environment to develop and test autonomy algorithms.

Current autonomous vehicle testing is done by industry using full-size vehicles on city streets – an expensive proposition not viable for the broader academic research community.
“There’s a barrier to entry for the science in the field,” said principal investigator James Rehg, a professor in the School of Interactive Computing. “Our platform uses a one-fifth scale vehicle, freeing us to do research at lower cost and without taking any risks – we can crash our car and it’s inexpensive to repair and nobody gets hurt.”

The autonomous cars will navigate the MiniCity and avoid hazards while obeying speed and traffic rules. Sensors will enable the cars to sense obstacles and make decisions on how fast to drive or how to steer. “We are addressing the issue of reproducibility of autonomous driving in a test environment,” Rehg said. 

Massachusetts Institute of Technology (MIT), one of TRI’s three original funded universities, is leading the research project. MIT operates an indoor autonomous driving track that simulates paved city streets. With Georgia Tech’s outdoor track, researchers can then see how autonomous cars perform over gravel, dirt, and other more realistic driving conditions. 

According to Rehg, autonomy testing presents unique challenges. “There’s a reason you get a driver’s test — you have to understand the variety of situations that can arise in driving and the rules, and you must understand how the context can change and make all the right decisions for safety.” 

With the MiniCity, Rehg and fellow investigator Evangelos Theodorou, an associate professor in the Daniel Guggenheim School of Aerospace Engineering, hope to develop a standardized testbed and protocol for testing and then invite academic teams to compete and measure the driving performance of their vehicles.  

Human-assist Robots to Help People Age in Place
Georgia Tech’s other TRI research project involves robotics that can assist older adults. It reflects Toyota and TRI’s priority to help older adults age in place.

“It’s really a powerful thing to have independence and be able to do things for yourself,” said the project’s principal investigator, Charlie Kemp, associate professor in the Wallace H. Coulter Department of Biomedical Engineering and adjunct associate professor in the School of Interactive Computing. Kemp also is a co-founder and the chief technology officer of Hello Robot Inc., a company that has commercialized robotic assistance technologies initially developed in his lab.  

Looking at the aging issue, Kemp and co-PI Hutchinson will examine how to take advantage of complementary characteristics that can lead to better physical collaboration between an individual and a robot.

“We are asking, ‘How can an individual and a particular robot best work together?’ ‘How do we individualize the robot to the person to give them a better quality of life?’” Kemp said.

They plan to take a modeling approach initially using physics simulations and, later, conducting studies with young able-bodied participants, healthy older adults, and older adults with impairments. 

The researchers will use sensing technology – including pressure sensors on beds that pinpoint a person’s body position and movement, as well as capacitive sensors that help the robot to better perceive a person’s body position up close. Such information can help with activities like dressing.  

“It’s a very intimate interaction between the robot and the human,” Hutchinson said.

Both investigators share TRI’s view that robotics that can assist older adults with daily living could make a major impact in the well-being of an increasingly graying population. In fact, during the next three decades, the global population over the age of 65 is projected to more than double. Japan, headquarters for Toyota, has the highest proportion of older citizens of any country in the world, with one in four people over 65.  

“We have talked about robots helping older adults for decades and we’re still not there,” said Kemp. “There’s a real opportunity to help people. As I get older, I’d love for this technology to be there for me and for my loved ones. While we still have a long way to go, the research can get us closer,” he added.

Hutchinson acknowledged that it will take time before people see robotic assistive technologies in hospitals or people’s homes, but the potential is there.

“What is most exciting about the TRI project is it has the potential to show up in people’s homes because TRI is invested in getting it there. And that means our research could really make an impact on a broad scale instead of only touching research journals or elite practitioners in the field,” he said.

Robotics Success Takes a Village 
The investigators agree that Georgia Tech’s multidisciplinary focus within robotics is a strength that will serve them well in their work with TRI, and especially in the future when autonomy goes mainstream.

“If you think about what it’s going to take for autonomous vehicles to really exist in the world on a large scale and deliver passengers in high volumes, it’s going to require all those things – engineering, science policy, law, and ethics – all those disciplines coming together,” said Rehg.
Kemp agreed, noting that since founding his Healthcare Robotics Lab in 2007, he’s attracted students from across engineering disciplines — from mechanical and computing to electrical, aerospace, and biomedical.  
“It's definitely something that's distinctive about Georgia Tech — it's a real strength,” he said.

Charlie Kemp owns equity in Hello Robot and is an inventor of Georgia Tech intellectual property (IP) licensed by Hello Robot. Consequently, he benefits from increases in the value of Hello Robot and receives royalties via Georgia Tech for sales made by Hello Robot. The terms of this arrangement have been reviewed and approved by Georgia Tech in accordance with its conflict-of-interest policies.

While other organisms form collective flocks, schools, or swarms for such purposes as mating, predation, and protection, the Lumbriculus variegatus worms are unusual in their ability to braid themselves together to accomplish tasks that unconnected individuals cannot. A new study reported by researchers at the Georgia Institute of Technology describes how the worms self-organize to act as entangled “active matter,” creating surprising collective behaviors whose principles have been applied to help blobs of simple robots evolve their own locomotion.

The research, supported by the National Science Foundation and the Army Research Office, was reported Feb. 5 in the journal Proceedings of the National Academy of Sciences. Findings from the work could help developers of swarm robots understand how emergent behavior of entangled active matter can produce unexpected, complex, and potentially useful mechanically driven behaviors.

Collective Behavior in Worms

The spark for the research came several years ago in California, where Saad Bhamla was intrigued by blobs of the worms he saw in a backyard pond.

“We were curious about why these worms would form these living blobs,” said Bhamla, an assistant professor in Georgia Tech’s School of Chemical and Biomolecular Engineering. “We have now shown through mathematical models and biological experiments that forming the blobs confers a kind of collective decision-making that enables worms in a larger blob to survive longer against desiccation. We also showed that they can move together, a collective behavior that’s not done by any other organisms we know of at the macro scale.”

Such collective behavior in living systems is of interest to researchers exploring ways to apply the principles of living systems to human-designed systems such as swarm robots, in which individuals must also work together to create complex behaviors.

“The worm blob collective turns out to have capabilities that are more than what the individuals have, a wonderful example of biological emergence,” said Daniel Goldman, a Dunn Family Professor in Georgia Tech’s School of Physics, who studies the physics of living systems.

Why the Worms Form Blobs

The worm blob system was studied extensively by Yasemin Ozkan-Aydin, a research associate in Goldman’s lab. Using bundles of worms she originally ordered from a California aquarium supply company – and now raises in Georgia Tech labs – Ozkan-Aydin put the worms through several experiments. Those included development of a “worm gymnasium” that allowed her to measure the strength of individual worms, knowledge important to understanding how small numbers of the creatures can move an entire blob.

She started by taking the aquatic worms out of the water and watching their behavior. First, they individually began searching for water. When that search failed, they formed a ball-shaped blob in which individuals took turns on the outer surface exposed to the air where evaporation was taking place – behavior she theorized would reduce the effect of evaporation on the collective. By studying the blobs, she learned that worms in a blob could survive out of water 10 times longer than individual worms could.

“They would certainly want to reduce desiccation, but the way in which they would do this is not obvious and points to a kind of collective intelligence in the system,” said Goldman. “They are not just surface-minimizing machines. They are looking to exploit good conditions and resources.”

Using Blobs to Escape Threats

Ozkan-Aydin also studied how worm blobs responded to both temperature gradients and intense light. The worms need a specific range of temperatures to survive and dislike intense light. When a blob was placed on a heated plate, it slowly moved away from the hotter portion of the plate to the cooler portion and under intense light formed tightly entangled blobs. The worms appeared to divide responsibilities for the movement, with some individuals pulling the blob while others helped lift the aggregation to reduce friction.

As with evaporation, the collective activity improves the chances of survival for the entire group, which can range from 10 worms up to as many as 50,000.

“For an individual worm going from hot to cold, survival depends on chance,” said Bhamla. “When they move as a blob, they move more slowly because they have to coordinate the mechanics. But if they move as a blob, 95% of them get to the cold side, so being part of the blob confers many survival advantages.”

A Worm Gymnasium

The researchers noted that only two or three “puller” worms were needed to drag a 15-worm blob. That led them to wonder just how strong the creatures were, so Ozkan-Aydin created a series of poles and cantilevers in which she could measure the forces exerted by individual worms. This “worm gymnasium” allowed her to appreciate how the pullers managed to do their jobs.

“When the worms are happy and cool, they stretch out and grab onto one of the poles with their heads and they pull onto it,” Bhamla said. “When they are pulling, you can see the deflection of the cantilever to which their tails were attached. Yasemin was able to use known weights to calibrate the forces the worms create. The force measurement shows the individual worms are packing a lot of power.”

Some worms were stronger than others, and as the temperature increased, their willingness to work out at the gym declined.

Applying Worm Principles to Robots

Ozkan-Aydin also applied the principles observed in the worms to small robotic blobs composed of “smart active particles,” six 3D-printed robots with two arms and two sensors allowing them to sense light. She added a mesh enclosure and pins to arms that allowed these “smarticles” to be entangled like the worms and tested a variety of gaits and movements that could be programmed into them.

“Depending on the intensity, the robots try to move away from the light,” Ozkan-Aydin said. “They generate emergent behavior that is similar to what we saw in the worms.”

She noted that there was no communication among the robots. “Each robot is doing its own thing in a decentralized way,” she said. “Using just the mechanical interaction and the attraction each robot had for light intensity, we could control the robot blob.”

By measuring the energy consumption of an individual robot when it performed different gaits (wiggle and crawl), she determined that the wiggle gait uses less power than the crawl gait. The researchers anticipate that by exploiting gait differentiation, future entangled robotic swarms could improve their energy efficiency. 

Expanding What Robot Swarms Can Do

The researchers hope to continue their study of the collective dynamics of the worm blobs and apply what they learn to swarm robots, which must work together with little communication to accomplish tasks that they could not do alone. But those systems must be able to work in the real world.

“Often people want to make robot swarms do specific things, but they tend to be operating in pristine environments with simple situations,” said Goldman. “With these blobs, the whole point is that they work only because of physical interaction among the individuals. That’s an interesting factor to bring into robotics.”

Among the challenges ahead are recruiting graduate students willing to work with the worm blobs, which have the consistency of bread dough. 

“The worms are very nice to work with,” said Ozkan-Aydin. “We can play with them and they are very friendly. But it takes a person who is very comfortable working with living systems.”

The project shows how the biological world can provide insights beneficial to the field of robotics, said Kathryn Dickson, program director of the Physiological Mechanisms and Biomechanics Program at the National Science Foundation.

“This discovery shows that observations of animal behavior in natural settings, along with biological experiments and modeling, can offer new insights, and how new knowledge gained from interdisciplinary research can help humans, for example, in the robotic control applications arising from this work,” she said.

This research was supported by the National Science Foundation (NSF) under grants CAREER 1941933 and 1817334 and the Army Research Office under grant W911NF-11-1-0514. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the sponsoring agencies.

CITATION: Yasemin Ozkan-Aydin, Daniel I. Goldman, and M. Saad Bhamla, “Collective dynamics in entangled worm and robot blobs. (Proceedings of the National Academy of Sciences, 2021). https://doi.org/10.1073/pnas.2010542118

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu)

Writer: John Toon

In the 19th century, scientists and engineers developed the discipline of statistical mechanics, which predicts how groups of simple particles transition between order and disorder, as when a collection of randomly colliding atoms freezes to form a uniform crystal lattice.

More challenging to predict are the collective behaviors that can be achieved when the particles become more complicated, such that they can move under their own power. This type of system — observed in bird flocks, bacterial colonies, and robot swarms — goes by the name "active matter.”

As reported in the January 1, 2021 issue of the journal Science, a team of physicists and engineers have proposed a new principle by which active matter systems can spontaneously order, without need for higher level instructions or even programmed interaction among the agents. And they have demonstrated this principle in a variety of systems, including groups of periodically shape-changing robots called "smarticles" — smart, active particles.

The theory, developed by Postdoctoral Researcher Pavel Chvykov at the Massachusetts Institute of Technology while a student of Prof. Jeremy England, who is now a researcher in the School of Physics at Georgia Institute of Technology, posits that certain types of active matter with sufficiently messy dynamics will spontaneously find what the researchers refer to as "low rattling" states.

“Rattling is when matter takes energy flowing into it and turns it into random motion,” England said. “Rattling can be greater either when the motion is more violent, or more random. Conversely, low rattling is either very slight or highly organized — or both. So, the idea is that if your matter and energy source allow for the possibility of a low rattling state, the system will randomly rearrange until it finds that state and then gets stuck there. If you supply energy through forces with a particular pattern, this means the selected state will discover a way for the matter to move that finely matches that pattern.”

To develop their theory, England and Chvykov took inspiration from a phenomenon — dubbed thermophoresis — discovered by the Swiss physicist Charles Soret in the late 19th century. In Soret's experiments, he discovered that subjecting an initially uniform salt solution in a tube to a difference in temperature would spontaneously lead to an increase in salt concentration in the colder region — which corresponds to an increase in order of the solution. 

Chvykov and England developed numerous mathematical models to demonstrate the low rattling principle, but it wasn't until they connected with Daniel Goldman, Dunn Family Professor of Physics at the Georgia Institute of Technology, that they were able to test their predictions. 

Said Goldman, "A few years back, I saw England give a seminar and thought that some of our smarticle robots might prove valuable to test this theory." Working with Chvykov, who visited Goldman's lab, Ph.D. students William Savoie and Akash Vardhan used three flapping smarticles enclosed in a ring to compare experiments to theory. The students observed that instead of displaying complicated dynamics and exploring the container completely, the robots would spontaneously self-organize into a few dances — for example, one dance consists of three robots slapping each other's arms in sequence. These dances could persist for hundreds of flaps, but suddenly lose stability and be replaced by a dance of a different pattern.

After first demonstrating that these simple dances were indeed low rattling states, Chvykov worked with engineers at Northwestern University, Prof. Todd Murphey and Ph.D. student Thomas Berrueta, who developed more refined and better controlled smarticles. The improved smarticles allowed the researchers to test the limits of the theory, including how the types and number of dances varied for different arm flapping patterns, as well as how these dances could be controlled. "By controlling sequences of low rattling states, we were able to make the system reach configurations that do useful work," Berrueta said. The Northwestern University researchers say that these findings may have broad practical implications for micro-robotic swarms, active matter, and metamaterials.

As England noted: “For robot swarms, it’s about getting many adaptive and smart group behaviors that you can design to be realized in a single swarm, even though the individual robots are relatively cheap and computationally simple. For living cells and novel materials, it might be about understanding what the ‘swarm’ of atoms or proteins can get you, as far as new material or computational properties.”

The study’s Georgia Tech-based team includes Jeremy L. England, a Physics of Living Systems scientist who researches with the School of Physics; Dunn Family Professor Daniel Goldman; professor Kurt Wiesenfeld, and graduate students Akash Vardhan (Quantitative Biosciences) and William Savoie (School of Physics). They join Pavel Chvykov (Massachusetts Institute of Technology), along with professor Todd D. Murphey and graduate students Thomas A. Berrueta and Alexander Samland of Northwestern University.

This material is based on work supported by the Army Research Office under awards from ARO W911NF-18-1-0101, ARO MURI Award W911NF-19-1-0233, ARO W911NF-13-1-0347, by the National Science Foundation under grants PoLS-0957659, PHY-1205878, PHY-1205878, PHY-1205878, and DMR-1551095, NSF CBET-1637764, by the James S. McDonnell Foundation Scholar Grant 220020476, and the Georgia Institute of Technology Dunn Family Professorship. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the sponsoring agencies.

CITATION: Chvykov & Berrueta, et al., “Low rattling: A predictive principle for self-organization in active collectives,” (Science 2021). https://science.sciencemag.org/content/371/6524/90/tab-pdf

The researchers demonstrated for the first time a multifunctional, magnetically responsive origami system, possessing distributed, untethered control capabilities. The untethered magnetic actuation separates the power source and controller out of the system, allowing scalable applications.

Researchers foresee that this actuation solution can be applied locally and remotely on complex origami assemblies. The actuation strategy enables a myriad of new applications, ranging from morphing robotics and satellites to biomedical devices.

“By distributively integrating the programmed magnetic soft materials into the bi-stable origami assembly, the magnetic actuation provides independent control of the folding and unfolding of each unit cell with instantaneous shape locking, which enables various robotic motion for functions such as tunable physical properties and configurable electronics for digital computing,” said principal investigator Ruike (Renee) Zhao, an assistant professor in the Department of Mechanical and Aerospace Engineering at Ohio State.

The research, "Untethered control of functional origami microrobots with distributed actuation," was reported Sept. 14 in the journal Proceedings of the National Academy of Sciences. The work was sponsored by the National Science Foundation (NSF).

Researchers have explored for decades how to leverage origami folding techniques in advanced engineering applications, such as morphing structures and devices. However, most actuation methods require physical bonds to external stimuli and lead to excessive wiring to provide the driving force for origami folding.

The new, untethered system is free from those rigid and often relatively bulky power sources, allowing faster speed and distributed actuation of the multifunctional structure.

To demonstrate this, researchers constructed a system of magnetic-responsive materials in a cylindrical origami pattern that consists of identical triangular panels known as a Kresling pattern. This pattern allows the cylinder’s walls to buckle under axial or torsional load.

“The Kresling pattern offers a very rich design space, which was crucial in coupling its mechanical response with magnetically responsive materials to achieve on-demand, untethered actuation, including our multifunctional origami for digital computing,” said Glaucio Paulino, professor and Raymond Allen Jones Chair in the Georgia Tech School of Civil and Environmental Engineering.

By controlling the magnetic field, researchers were able to control the direction, intensity, and speed of the material’s folding and deployment. In the tests, researchers achieved untethered actuation as fast as one tenth of a second with instantaneous shape locking.

Next, researchers attached a magnetized plate to each of the Kresling unit cells. This allowed them to utilize a two-dimensional magnetic field to actuate the unit cells simultaneously or independently by using different magnetic torques of the plates and distinct geometric-mechanical properties of each unit cell.

“The multi-unit Kresling assembly is an origami robot in which the bi-stable folding and unfolding create robotic motion. It can passively sense and actively respond to the external environment. By integrating electronic circuits into the origami robot, it further enables intelligent autonomous robots with integrated actuation, sensing, and decision making,” Zhao said. “For example, the external pressure or forces that act on the robot will trigger the passive folding of the robot, indicating the presence of an obstacle. The robot can then actively unfold itself and decide the next move.”

The untethered magnetic control pushes the boundary of the application of origami systems, which could lead to solutions of next-generation biomimetic soft robots and robotic systems for advanced engineering applications.

“We anticipate that the reported magnetic origami system is applicable beyond the bounds of this work, including future origami-inspired robots, morphing mechanisms, biomedical devices, and outer space structures,” Paulino said. 

This research was supported by Prof. Zhao’s two recent NSF Awards from the Mechanics of Materials and Structures program (NSF Award #1943070, #1939543) and Ohio State’s Institute of Material Research. The authors at Georgia Tech acknowledge NSF (Award #1538830) and the Raymond Allen Jones Chair. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

- Written by The Ohio State University

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu).

Built by robotics engineers at the Georgia Institute of Technology to take advantage of the low-energy lifestyle of real sloths, SlothBot demonstrates how being slow can be ideal for certain applications. Powered by solar panels and using innovative power management technology, SlothBot moves along a cable strung between two large trees as it monitors temperature, weather, carbon dioxide levels, and other information in the Garden’s 30-acre midtown Atlanta forest.

“SlothBot embraces slowness as a design principle,” said Magnus Egerstedt, professor and Steve W. Chaddick School Chair in the Georgia Tech School of Electrical and Computer Engineering. “That’s not how robots are typically designed today, but being slow and hyper-energy efficient will allow SlothBot to linger in the environment to observe things we can only see by being present continuously for months, or even years.”

About three feet long, SlothBot’s whimsical 3D-printed shell helps protect its motors, gearing, batteries, and sensing equipment from the weather. The robot is programmed to move only when necessary, and will locate sunlight when its batteries need recharging. At the Atlanta Botanical Garden, SlothBot will operate on a single 100-foot cable, but in larger environmental applications, it will be able to switch from cable to cable to cover more territory.

“The most exciting goal we’ll demonstrate with SlothBot is the union of robotics and technology with conservation,” said Emily Coffey, vice president for conservation and research at the Garden. “We do conservation research on imperiled plants and ecosystems around the world, and SlothBot will help us find new and exciting ways to advance our research and conservation goals.”

Supported by the National Science Foundation and the Office of Naval Research, SlothBot could help scientists better understand the abiotic factors affecting critical ecosystems, providing a new tool for developing information needed to protect rare species and endangered ecosystems.

“SlothBot could do some of our research remotely and help us understand what’s happening with pollinators, interactions between plants and animals, and other phenomena that are difficult to observe otherwise,” Coffey added. “With the rapid loss of biodiversity and with more than a quarter of the world’s plants potentially heading toward extinction, SlothBot offers us another way to work toward conserving those species.”

Inspiration for the robot came from a visit Egerstedt made to a vineyard in Costa Rica where he saw two-toed sloths creeping along overhead wires in their search for food in the tree canopy. “It turns out that they were strategically slow, which is what we need if we want to deploy robots for long periods of time,” he said.

A few other robotic systems have already demonstrated the value of slowness. Among the best known are the Mars Exploration Rovers that gathered information on the red planet for more than a dozen years. “Speed wasn’t really all that important to the Mars Rovers,” Egerstedt noted. “But they learned a lot during their leisurely exploration of the planet.”

Beyond conservation, SlothBot could have applications for precision agriculture, where the robot’s camera and other sensors traveling in overhead wires could provide early detection of crop diseases, measure humidity, and watch for insect infestation. After testing in the Atlanta Botanical Garden, the researchers hope to move SlothBot to South America to observe orchid pollination or the lives of endangered frogs.

The research team, which includes Ph.D students Gennaro Notomista and Yousef Emam, undergraduate student Amy Yao, and postdoctoral researcher Sean Wilson, considered multiple locomotion techniques for the SlothBot. Wheeled robots are common, but in the natural world they can easily be defeated by obstacles like rocks or mud. Flying robots require too much energy to linger for long. That’s why Egerstedt’s observation of the wire-crawling sloths was so important.

“It’s really fascinating to think about robots becoming part of the environment, a member of an ecosystem,” he said. “While we’re not building an anatomical replica of the living sloth, we believe our robot can be integrated to be part of the ecosystem it’s observing like a real sloth.”

The SlothBot launched in the Atlanta Botanical Garden is the second version of a system originally reported in May 2019 at the International Conference on Robotics and Automation. That robot was a much smaller laboratory prototype.

Beyond their conservation goals, the researchers hope SlothBot will provide a new way to stimulate interest in conservation from the Garden’s visitors. “This will help us tell the story of the merger between technology and conservation,” Coffey said. “It’s a unique way to engage the public and bring forward a new way to tell our story.”

And that should be especially interesting to children visiting the Garden.

“This new way of thinking about robots should trigger curiosity among the kids who will walk by it,” said Egerstedt. “Thanks to SlothBot, I’m hoping we will get an entirely new generation interested in what robotics can do to make the world better.”

This research was sponsored by the U.S. Office of Naval Research through Grant N00014-15-2115 and by the National Science Foundation through Grant 1531195. The content is solely the responsibility of the authors and does not necessarily represent the official views of the sponsoring agencies.

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contacts: John Toon, Georgia Tech (404-894-6986) (jtoon@gatech.edu); Danny Flanders, Atlanta Botanical Garden (404-591-1550) (dflanders@atlantabg.org).

Writer: John Toon

The studies were originally intended to test for gender bias, that is, if people thought a robot believed to be female may be less competent at some jobs than a robot believed to be male and vice versa. The studies' titles even included the words "gender," "stereotypes," and "preference," but researchers at the Georgia Institute of Technology discovered no significant sexism against the machines.

“This did surprise us. There was only a very slight difference in a couple of jobs but not significant. There was, for example, a small preference for a male robot over a female robot as a package deliverer,” said Ayanna Howard, the principal investigator in both studies. Howard is a professor in and the chair of Georgia Tech’s School of Interactive Computing.

Although robots are not sentient, as people increasingly interface with them, we begin to humanize the machines. Howard studies what goes right as we integrate robots into society and what goes wrong, and much of both has to do with how the humans feel around robots.

I hate robots

“Surveillance robots are not socially engaging, but when we see them, we still may act like we would when we see a police officer, maybe not jaywalking and being very conscientious of our behavior,” said Howard, who is also Linda J. and Mark C. Smith Chair and Professor in Bioengineering in Georgia Tech’s School of Electrical and Computer Engineering.

“Then there are emotionally engaging robots designed to tap into our feelings and work with our behavior. If you look at these examples, they lead us to treat these robots as if they were fellow intelligent beings.”

It’s a good thing robots don’t have feelings because what study participants lacked in gender bias they more than made up for in judgments against the humanoid robots' competence. That predisposition was so strong that Howard wondered if it may have overridden any potential gender biases against robots – after all, social science studies have shown that gender biases are still prevalent with respect to human jobs, even if implicit.

In questionnaires, humanoid robots introduced themselves via video to randomly recruited online survey respondents, who ranged in age from their twenties to their seventies and were mostly college-educated. The humans ranked robots’ career competencies compared to human abilities, only trusting the machines to competently perform a handful of simple jobs. 

Pass the scalpel

“The results baffled us because the things that people thought robots were less able to do were things that they do well. One was the profession of surgeon. There are Da Vinci robots that are pervasive in surgical suites, but respondents didn’t think robots were competent enough,” Howard said. “Security guard – people didn’t think robots were competent at that, and there are companies that specialize in great robot security.”

Cumulatively, the 200 participants across the two studies thought robots would also fail as nannies, therapists, nurses, firefighters, and totally bomb as comedians. But they felt confident bots would make fantastic package deliverers and receptionists, pretty good servers, and solid tour guides.

The researchers could not say where the competence biases originate. Howard could only speculate that some of the bad rap may have come from media stories of robots doing things like falling into swimming pools or injuring people.

It’s a boy 

Despite the lack of gender bias, participants readily assigned genders to the humanoid robots. For example, people accepted gender prompts by robots introducing themselves in videos.

If a robot said, “Hi, my name is James,” in a male-sounding voice, people mostly identified the robot as male. If it said, “Hi, my name is Mary,” in a female voice, people mostly said it was female.

Some robots greeted people by saying “Hi” in a neutral sounding voice, and still, most participants assigned the robot a gender. The most common choice was male followed by neutral then by female. For Howard, this was an important takeaway from the study for robot developers.

“Developers should not force gender on robots. People are going to gender according to their own experiences. Give the user that right. Don’t reinforce gender stereotypes,” Howard said.

Social is good

Some in the field advocate for not building robots in humanoid form at all in order to discourage any kind of humanization, but the Georgia Tech team takes a less stringent approach.

"There is no single one-size-fits-all answer on whether it is appropriate to design robots to look like human beings.  It depends on a variety of ethical considerations and other factors, including whether people might trust a robot too much if it has a human-like appearance," said Jason Borenstein, a co-principal investigator on one of the papers and an ethics researcher in Georgia Tech's School of Public Policy.

“Robots can be good for social interaction. They could be very helpful in elder care facilities to keep people company. They might also make better nannies than letting the TV babysit the kids,” said Howard, who also defended robots’ comedic talent, provided they are programmed for that.

“If you ever go to an amusement park, there are animatronics that tell really good jokes.”

Read the studies

The two studies were submitted to conferences that were canceled due to COVID-19.

Why Should We Gender? The Effect of Robot Gendering and Occupational Stereotypes on Human Trust and Perceived Competency was published in Proceedings of 2020 ACM Conference on Human-Robot Interaction (HRI’20), which appeared in March 2020. Robot Gendering: Influences on Trust, Occupational Competency, and Preference of Robot Over Human appeared in CHI 2020 Extended Abstracts (computer-human interaction, DOI: 10.1145/3334480.3382930).

The research was funded by the National Science Foundation and by the Alfred P. Sloan Foundation.

The papers’ coauthors were De’Aira Bryant, Kantwon Rogers, and Jason Borenstein from Georgia Tech. The National Science foundation funded via grant 1849101. The Alfred P. Sloan Foundation funded via grant G-2019-11435. Any findings, conclusions, or recommendations are those of the authors and not necessarily of the sponsors.

Also read: Surfaces that grip like gecko feet may come to an assembly line near you

Here's how to subscribe to our free science and technology email newsletter

Writer & media inquiries: Ben Brumfield (404-272-2780), email: ben.brumfield@comm.gatech.edu

Georgia Institute of Technology

Built with wheeled appendages that can be lifted and wheels able to wiggle, a new robot known as the “Mini Rover” has developed and tested complex locomotion techniques robust enough to help it climb hills covered with such granular material – and avoid the risk of getting ignominiously stuck on some remote planet or moon. 

Using a complex move the researchers dubbed “rear rotator pedaling,” the robot can climb a slope by using its unique design to combine paddling, walking, and wheel spinning motions. The rover’s behaviors were modeled using a branch of physics known as terradynamics.

“When loose materials flow, that can create problems for robots moving across it,” said Dan Goldman, the Dunn Family Professor in the School of Physics at the Georgia Institute of Technology. “This rover has enough degrees of freedom that it can get out of jams pretty effectively. By avalanching materials from the front wheels, it creates a localized fluid hill for the back wheels that is not as steep as the real slope. The rover is always self-generating and self-organizing a good hill for itself.”

The research was reported on May 13 as the cover article in the journal Science Robotics. The work was supported by the NASA National Robotics Initiative and the Army Research Office.

A robot built by NASA’s Johnson Space Center pioneered the ability to spin its wheels, sweep the surface with those wheels and lift each of its wheeled appendages where necessary, creating a broad range of potential motions. Using in-house 3D printers, the Georgia Tech researchers collaborated with the Johnson Space Center to re-create those capabilities in a scaled-down vehicle with four wheeled appendages driven by 12 different motors.

“The rover was developed with a modular mechatronic architecture, commercially available components, and a minimal number of parts,” said Siddharth Shrivastava, an undergraduate student in Georgia Tech’s George W. Woodruff School of Mechanical Engineering. “This enabled our team to use our robot as a robust laboratory tool and focus our efforts on exploring creative and interesting experiments without worrying about damaging the rover, service downtime, or hitting performance limitations.” 

The rover’s broad range of movements gave the research team an opportunity to test many variations that were studied using granular drag force measurements and modified Resistive Force Theory. Shrivastava and School of Physics Ph.D. candidate Andras Karsai began with the gaits explored by the NASA RP15 robot, and were able to experiment with locomotion schemes that could not have been tested on a full-size rover.

The researchers also tested their experimental gaits on slopes designed to simulate planetary and lunar hills using a fluidized bed system known as SCATTER (Systematic Creation of Arbitrary Terrain and Testing of Exploratory Robots) that could be tilted to evaluate the role of controlling the granular substrate. Karsai and Shrivastava collaborated with Yasemin Ozkan-Aydin, a postdoctoral research fellow in Goldman’s lab, to study the rover motion in the SCATTER test facility. 

“By creating a small robot with capabilities similar to the RP15 rover, we could test the principles of locomoting with various gaits in a controlled laboratory environment,” Karsai said. “In our tests, we primarily varied the gait, the locomotion medium, and the slope the robot had to climb. We quickly iterated over many gait strategies and terrain conditions to examine the phenomena that emerged.”

In the paper, the authors describe a gait that allowed the rover to climb a steep slope with the front wheels stirring up the granular material – poppy seeds for the lab testing – and pushing them back toward the rear wheels. The rear wheels wiggled from side-to-side, lifting and spinning to create a motion that resembles paddling in water. The material pushed to the back wheels effectively changed the slope the rear wheels had to climb, allowing the rover to make steady progress up a hill that might have stopped a simple wheeled robot.

The experiments provided a variation on earlier robophysics work in Goldman’s group that involved moving with legs or flippers, which had emphasized disturbing the granular surfaces as little as possible to avoid getting the robot stuck.

“In our previous studies of pure legged robots, modeled on animals, we had kind of figured out that the secret was to not make a mess,” said Goldman. “If you end up making too much of a mess with most robots, you end up just paddling and digging into the granular material. If you want fast locomotion, we found that you should try to keep the material as solid as possible by tweaking the parameters of motion.”

But simple motions had proved problematic for Mars rovers, which got stuck in granular materials. Goldman says the gait discovered by Shrivastava, Karsai and Ozkan-Aydin might be able to help future rovers avoid that fate.

“This combination of lifting and wheeling and paddling, if used properly, provides the ability to maintain some forward progress even if it is slow,” Goldman said. “Through our laboratory experiments, we have shown principles that could lead to improved robustness in planetary exploration – and even in challenging surfaces on our own planet.”

The researchers hope next to scale up the unusual gaits to larger robots, and to explore the idea of studying robots and their localized environments together. “We’d like to think about the locomotor and its environment as a single entity,” Goldman said. “There are certainly some interesting granular and soft matter physics issues to explore.”

Though the Mini Rover was designed to study lunar and planetary exploration, the lessons learned could also be applicable to terrestrial locomotion – an area of interest to the Army Research Laboratory, one of the project’s sponsors.

"This basic research is revealing exciting new approaches for locomotion in complex terrain," said Dr. Samuel Stanton, program manager, Army Research Office, an element of the U.S. Army Combat Capabilities Development Command's Army Research Laboratory. "This could lead to platforms capable of intelligently transitioning between wheeled and legged modes of movement to maintain high operational tempo."

Beyond those already mentioned, the researchers worked with Robert Ambrose and William Bluethmann at NASA, and traveled to NASA JSC to study the full-size NASA RP15 rover.

This work was supported by the Army Research Office (W911NF-18-1-0120) and the NASA National Robotics Initiative (NNX15AR21G). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the sponsoring agencies.

CITATION: Siddharth Shrivastava, Andras Karsai, Yasemin Ozkan-Aydin, Ross Pettinger, William Bluethmann, Robert O. Ambrose, Daniel I. Goldman, “Material remodeling on granular terrain yields robustness benefits for a robophysical rover.” (Science Robotics, May 2020)

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu)

Writer: John Toon

The images, taken by a robotic underwater vehicle, were part of a broad set of data collected in a variety of experiments by an international team. The International Thwaites Glacier Collaboration (ITGC) announced the completion of this first-ever major research venture on the glacier coincident with the 200-year anniversary of the discovery of Antarctica in 1820.

Already, Thwaites accounts for about four percent of global sea-level rise. Researchers have had concerns that a tipping point in the stability at its foundations could result in a run-away collapse of the glacier and boost sea levels by as much as 25 inches. By studying multiple aspects of Thwaites, the ITGC wants to understand more about the likelihood that the glacier the size of Florida may reach such instability in the coming decades.

Line of concern

The area of concern that the underwater vehicle visited is called the grounding line, and it is important to the stability of Thwaites Glacier’s footing. It is the line between where the glacier rests on the ocean bed and where it floats over water. The farther back the grounding line recedes, the faster the ice can flow into the sea, pushing up sea-level. 

“Visiting the grounding line is one of the reasons work like this is important because we can drive right up to it and actually measure where it is,” said Britney Schmidt, an ITGC co-investigator from the Georgia Institute of Technology. “It's the first time anyone has done that or has ever even seen the grounding zone of a major glacier under the water, and that’s the place where the greatest degree of melting and destabilization can occur.”

The underwater robot, Icefin, was engineered by Schmidt’s Georgia Tech lab. The Georgia Tech team was part of a greater collaboration between researchers from the U.S. and the British Antarctic Survey (BAS), who lived and worked on Thwaites in December and January. A BAS hot water drill melted a hole 590 meters deep (1,935 feet) to access the ocean cavity for Icefin.

“Icefin swam over 15 km (9.3 miles) round trip during five missions. This included two passes up to the grounding zone, including one where we got as close as we physically could to the place where the seafloor meets the ice,” said Schmidt, who is an associate professor in Georgia Tech’s School of Earth and Atmospheric Sciences. “We saw amazing ice interactions driven by sediments at the line and from the rapid melting from warm ocean water.”

^{[Ready for graduate school? Here's how to apply to Georgia Tech.]} 

Historic research venture

In the coming months and years, the ITGC team made up of researchers from multiple universities and research institutions in the U.S. and the UK will publish studies with thorough findings based on the unprecedented data collected during the field campaign.

The array of research the scientists carried out research included seismic and radar measurements and using hot water drills to make holes between 300 and 700 meters (985 and 2,300 feet) deep down to the ocean and glacier bed below Thwaites’ ice. Researchers also took cores of sediment from the seafloor and under parts of the glacier grounded on the bed to examine the quality of the foothold that it offers Thwaites.

“We know that warmer ocean waters are eroding many of West Antarctica’s glaciers, but we’re particularly concerned about Thwaites. This new data will provide a new perspective of the processes taking place, so we can predict future change with more certainty,” said Keith Nicholls, an oceanographer from the British Antarctic Survey.

Nicholls is a co-principal investigator on the project that involved Schmidt along with David Holland of New York University. The research is funded by the National Science Foundation, the UK Natural Environment Research Council, the U.S. Antarctic Program, and the British Antarctic Survey.

Antarctica sea-level background

Over the past 30 years, the amount of ice flowing to the sea from Thwaites and its neighboring glaciers has nearly doubled.

“While Greenland's contribution to sea level has already reached an alarming rate, Antarctica is just now picking up its contributions to sea level,” Schmidt said. “It has the largest body of ice on Earth and will contribute more and more of sea-level rise over the next 100 years and beyond. It’s a massive source of uncertainty in the climate system.”

Watch BBC News report on this research.

External News Coverage: 

BBC News- Antarctica melting: Climate change and the journey to the 'doomsday glacier' 

The Atlantic- The New Video of One of the Scariest Places on Earth

The Washington Post- Unprecedented data confirms that Antarctica’s most dangerous glacier is melting from below

BBC Newsround- Climate change: Scientists concerned about future of Antarctic glacier

Daily Mail Online- Scientists drill into Antarctica's 'doomsday' Thwaites glacier for the first time in a bid to stop dramatic sea level rise as the ice shelf the size of BRITAIN melts at an alarming rate

Yahoo News- Scientists drill into ‘doomsday glacier’ the size of Britain to see if it’s going to collapse

Fox News- Antarctica’s ‘doomsday glacier’ reveals alarming new trait to scientists

Cosmos Magazine- Here's what's below an unstable glacier

PBS Newshour- A risky expedition to study the ‘doomsday glacier’ 

NOVA Next- Scientists find warm water beneath Antarctica’s most at-risk glacier

More reading: Instability in Antarctic Ice Projected to Make Sea Level Rise Rapidly and

Reframing Antarctica’s Meltwater Pond Dangers to Ice Shelves and Sea Level

Any findings, conclusions, or recommendations are those of the authors and not necessarily of the sponsors.

Writer & Media Representative: Ben Brumfield (404-272-2780), email: ben.brumfield@comm.gatech.edu

Georgia Institute of Technology

To help young people struggling with those conditions – or orthopedic problems like clubfoot, scoliosis, and osteogenesis imperfecta, among other things – Shriners Hospitals for Children® and the Georgia Institute of Technology have launched an ambitious collaborative research effort to address these conditions, including the development of devices to facilitate limb movement and function.

The new research affiliation brings together the clinical, surgical, and scientific expertise of Shriners Hospitals for Children physicians and researchers with Georgia Tech’s cutting-edge expertise in biomedical engineering, robotics, and device development. The coordinated effort also will leverage the two organizations’ proficiency in big data and artificial intelligence tools for personalized medicine, according to Marc Lalande, Ph.D., vice president of research programs for Shriners Hospitals for Children.

“Our joint goals, through genetic and genomic data gathered by Shriners Hospitals for Children, are to improve patient therapeutic responses by optimizing individualized treatment regimens and reducing adverse events,” Lalande said.

Several joint projects already are underway.

Jaydev Desai, professor in the Wallace H. Coulter Department of Biomedical Engineering (BME) at Georgia Tech and Emory University, is working with Scott Kozin, M.D., chief of staff and hand surgeon at Shriners Hospitals for Children-Philadelphia, on a wearable customized robotic exoskeleton with voice recognition for children with cervical spine injury.

“This is a patient specific system for kids with spinal cord injury,” explained Desai, who is director of the Georgia Center for Medical Robotics and associate director of Georgia Tech’s Institute for Robotics and Intelligent Machines. “The system is designed to translate voice commands into actions, meaning the exoskeleton will conform to the proper shape and posture of the fingers, so to speak, depending on the task. The idea is to enhance the child’s ability to perform the activities of daily living.”

Kozin expects his patients with spinal cord injuries will benefit from Georgia Tech’s innovative pediatric prosthesis development – its utility, actuation, and dexterity. “Alternative pathways for the recovery of sensation will enhance their function and independence. We are excited about this new collaboration combining institutions with similar missions and visions devoted to improving the lives of children,” said Kozin, who also is collaborating with Georgia Tech’s Frank Hammond (assistant professor in BME and mechanical engineering) on wearable sensory transfer devices for patients with diminished peripheral sensation or amputations, improving their ability to use intuitively powered prostheses and orthoses. 

Additionally, Aaron Young, assistant professor in the George W. Woodruff School of Mechanical Engineering at Georgia Tech, is working with David Westberry, M.D., pediatric orthopedic surgeon at Shriners Hospitals for Children-Greenville, on a smart robotic exoskeleton designed to address excessive knee flexion (crouch gait), a condition common in patients with cerebral palsy. The condition can lead to permanent joint deformity if untreated, as well as reduced independence and locomotion capability.

“The device is basically a lightweight, wearable robot designed to assist physical therapists working on pediatric mobility – the idea is to essentially retrain the child’s neuroplasticity,” said Young, who is testing the device with Westberry at Shriners Hospitals for Children-Greenville in South Carolina. “The exciting thing about Shriners Hospitals for Children-Greenville is that it has an advanced motion analysis center where Shriners’ physicians and researchers are looking at not just the child’s gait, but also at the internal mechanics. It’s very rewarding to collaborate with the Shriners team – they are very quantitative in their approach to treatment.”

That quantitative approach includes the integration of biomedical informatics, data science, and artificial intelligence into the clinical research programs of the Shriners Hospitals for Children network of 14 pediatric motion analysis centers and the healthcare system’s newly launched Genomics Institute. As part of this process, researchers are collaborating with Dongmei Wang, BME professor at Georgia Tech, where she is director of the Biomedical Informatics and Bioimaging Lab.

“This collaboration is extremely important for us because not only have we committed to work on a major national need in youth health, but also because we have been planning to establish a pediatric big data center using advanced IT and AI,” said Wang, whose collaborators at Shriners Hospitals for Children include Gerald Harris (Motion Analysis, Shriners Hospitals for Children-Chicago) and Kamran Shazand (Shriners Hospitals for Children Genomics Institute, Tampa, Florida). 

“Our lab has piloted multiple pediatric projects,” Wang said. “But this project represents a quantum leap, taking our work to the next level, in a real-world pediatric care setting. Shriners Hospitals for Children is a perfect fit for us.”

Leanne West, Georgia Tech’s chief engineer of pediatric technologies, said she’s looking forward to “the unique research opportunities this relationship with Shriners Hospitals for Children will provide. It will be exciting to see what is possible for us to achieve together.”

About pediatric device research at Georgia Tech
Georgia Tech’s wide-ranging efforts in pediatric device development brings the institute’s engineers and scientists together with clinical experts and researchers to develop innovative technological solutions to problems in the health and care of children. The work provides opportunities for interdisciplinary collaboration in pediatrics, creating breakthrough discoveries, enhancing the lives of children and young adults.

About Shriners Hospitals for Children 
Shriners Hospitals for Children is changing lives every day through innovative pediatric specialty care, world-class research, and outstanding medical education. Its healthcare system provides care for children with orthopedic conditions, burns, spinal cord injuries, and cleft lip and palate. All care and services are provided regardless of families’ ability to pay. Since opening its first location in 1922, the healthcare system has treated more than 1.4 million children. For more information, visit shrinershospitalsforchildren.org.

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu).

Writer: Jerry Grillo

Meet the HoneyBot. 

Developed by a team of researchers at the Georgia Institute of Technology, the diminutive device is designed to lure in digital troublemakers who have set their sights on industrial facilities. HoneyBot will then trick the bad actors into giving up valuable information to cybersecurity professionals.

The decoy robot arrives as more and more devices – never designed to operate on the Internet – are coming online in homes and factories alike, opening up a new range of possibilities for hackers looking to wreak havoc in both the digital and physical world.

“Robots do more now than they ever have, and some companies are moving forward with, not just the assembly line robots, but free-standing robots that can actually drive around factory floors,” said Raheem Beyah, the Motorola Foundation Professor and interim Steve W. Chaddick School Chair in Georgia Tech’s School of Electrical and Computer Engineering. “In that type of setting, you can imagine how dangerous this could be if a hacker gains access to those machines. At a minimum, they could cause harm to whatever products are being produced. If it’s a large enough robot, it could destroy parts or the assembly line. In a worst-case scenario, it could injure or cause death to the humans in the vicinity.”

Internet security professionals long have employed decoy computer systems known as “honeypots” as a way to throw cyberattackers off the trail. The research team applied the same concept to the HoneyBot, which is partially funded with a grant from the National Science Foundation. Once hackers gain access to the decoy, they leave behind valuable information that can help companies further secure their networks.

“A lot of cyberattacks go unanswered or unpunished because there’s this level of anonymity afforded to malicious actors on the internet, and it’s hard for companies to say who is responsible,” said Celine Irvene, a Georgia Tech graduate student who worked with Beyah to devise the new robot. “Honeypots give security professionals the ability to study the attackers, determine what methods they are using, and figure out where they are or potentially even who they are.”

The gadget can be monitored and controlled through the internet. But unlike other remote-controlled robots, the HoneyBot’s special ability is tricking its operators into thinking it is performing one task, when in reality it’s doing something completely different.

“The idea behind a honeypot is that you don’t want the attackers to know they’re in a honeypot,” Beyah said. “If the attacker is smart and is looking out for the potential of a honeypot, maybe they’d look at different sensors on the robot, like an accelerometer or speedometer, to verify the robot is doing what it had been instructed. That’s where we would be spoofing that information as well. The hacker would see from looking at the sensors that acceleration occurred from point A to point B.”

In a factory setting, such a HoneyBot robot could sit motionless in a corner, springing to life when a hacker gains access – a visual indicator that a malicious actor is targeting the facility.

Rather than allowing the hacker to then run amok in the physical world, the robot could be designed to follow certain commands deemed harmless – such as meandering slowly about or picking up objects – but stopping short of actually doing anything dangerous.

So far, their technique seems to be working.

In experiments designed to test how convincing the false sensor data would be to individuals remotely controlling the device, volunteers in December 2017 used a virtual interface to control the robot and could not to see what was happening in real life. To entice the volunteers to break the rules, at specific spots within the maze, they encountered forbidden “shortcuts” that would allow them to finish the maze faster.

In the real maze back in the lab, no shortcut existed, and if the participants opted to go through it, the robot instead remained still. Meanwhile, the volunteers – who have now unwittingly become hackers for the purposes of the experiment – were fed simulated sensor data indicating they passed through the shortcut and continued along.

“We wanted to make sure they felt that this robot was doing this real thing,” Beyah said.

In surveys after the experiment, participants who actually controlled the device the whole time and those who were being fed simulated data about the fake shortcut both indicated that the data was believable at similar rates.

“This is a good sign because it indicates that we’re on the right track,” Irvene said.

This material is based upon work supported by the National Science Foundation under Grant No. 1544332. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

The 3D-printed smarticles — short for smart active particles — can do just one thing: flap their two arms. But when five of these smarticles are confined in a circle, they begin to nudge one another, forming a robophysical system known as a “supersmarticle” that can move by itself. Adding a light or sound sensor allows the supersmarticle to move in response to the stimulus — and even be controlled well enough to navigate a maze.

Though rudimentary now, the notion of making robots from smaller robots — and taking advantage of the group capabilities that arise by combining individuals — could provide mechanically based control over very small robots. Ultimately, the emergent behavior of the group could provide a new locomotion and control approach for small robots that could potentially change shapes.

“These are very rudimentary robots whose behavior is dominated by mechanics and the laws of physics,” said Dan Goldman, a Dunn Family Professor in the School of Physics at the Georgia Institute of Technology. “We are not looking to put sophisticated control, sensing, and computation on them all. As robots become smaller and smaller, we’ll have to use mechanics and physics principles to control them because they won’t have the level of computation and sensing we would need for conventional control.”

The research, which was supported by the Army Research Office and the National Science Foundation, was reported September 18 in the journal Science Robotics. Researchers from Northwestern University also contributed to the project.

The foundation for the research came from an unlikely source: a study of construction staples. By pouring these heavy-duty staples into a container with removable sides, former Ph.D. student Nick Gravish — now a faculty member at the University of California San Diego — created structures that would stand by themselves after the container’s walls were removed.

Shaking the staple towers eventually caused them to collapse, but the observations led to a realization that simple entangling of mechanical objects could create structures with capabilities well beyond those of the individual components. 

“A robot made of other rudimentary robots became the vision,” Goldman said. “You could imagine making a robot in which you would tweak its geometric parameters a bit and what emerges is qualitatively new behaviors.”

To explore the concept, graduate research assistant Will Savoie used a 3D printer to create battery-powered smarticles, which have motors, simple sensors, and limited computing power. The devices can change their location only when they interact with other devices while enclosed by a ring.

“Even though no individual robot could move on its own, the cloud composed of multiple robots could move as it pushed itself apart and shrink as it pulled itself together,” Goldman explained. “If you put a ring around the cloud of little robots, they start kicking each other around, and the larger ring — what we call a supersmarticle — moves around randomly.”

The researchers noticed that if one small robot stopped moving, perhaps because its battery died, the group of smarticles would begin moving in the direction of that stalled robot. Graduate student Ross Warkentin learned he could control the movement by adding photo sensors to the robots that halt the arm flapping when a strong beam of light hits one of them.

“If you angle the flashlight just right, you can highlight the robot you want to be inactive, and that causes the ring to lurch toward or away from it, even though no robots are programmed to move toward the light,” Goldman said. “That allowed steering of the ensemble in a very rudimentary, stochastic way.”

School of Physics Professor Kurt Wiesenfeld and graduate student Zack Jackson modeled the movement of the these smarticles and supersmarticles to understand how the nudges and mass of the ring affected overall movement. Researchers from Northwestern University studied how the interactions between the smarticles provided directional control.

"For many robots, we have electrical current move motors that generate forces on parts that collectively move a robot reliably,” said Todd Murphey, a professor of mechanical engineering who worked with Northwestern graduate students Thomas Berrueta and Ana Pervan. “We learned that although individual smarticles interact with each other through a chaos of wiggling impacts that are each unpredictable, the whole robot composed of those smarticles moves predictably and in a way that we can exploit in software."

In future work, Goldman envisions more complex interactions that utilize the simple sensing and movement capabilities of the smarticles. “People have been interested in making a certain kind of swarm robots that are composed of other robots,” he said. “These structures could be reconfigured on demand to meet specific needs by tweaking their geometry.”

The project is of interest to the U.S. Army because it could lead to new robotic systems capable of changing their shapes, modalities and functions, said Sam Stanton. He is program manager of complex dynamics and systems at the Army Research Office, an element of U.S. Army Combat Capabilities Development Command’s Army Research Laboratory.

“Future Army unmanned systems and networks of systems are imagined to be capable of transforming their shape, modality, and function. For example, a robotic swarm may someday be capable of moving to a river and then autonomously forming a structure to span the gap,” Stanton said. “Dan Goldman's research is identifying physical principles that may prove essential for engineering emergent behavior in future robot collectives as well as new understanding of fundamental tradeoffs in system performance, responsiveness, uncertainty, resiliency, and adaptivity.”

In addition to those already mentioned, the research also included Georgia Tech graduate student Shengkai Li.

This material is based upon work supported by the Army Research Office under award W911NF-13-1-0347 and by the National Science Foundation under grants PoLS-0957659, PHY-1205878, DMR-1551095, PHY-1205878. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the sponsoring agencies.

CITATION: William Savoie, et al., “A robot made of robots: emergent transport and control of a smarticle ensemble,” (Science Robotics 2019). 

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Assistance: John Toon (404-894-6986) (jtoon@gatech.edu).

Writer: John Toon

Normally, tedious modeling of mechanics, electronics, and information science is required to understand how insects’ or robots’ moving parts coordinate smoothly to take them places. But in a new study, biomechanics researchers at the Georgia Institute of Technology boiled down the sprints of cockroaches to handy principles and equations they then used to make a test robot amble about better.

The method told the researchers about how each leg operates on its own, how they all come together as a whole, and the harmony or lack thereof in how they do it. Despite bugs’ and bots’ utterly divergent motion dynamics, the new method worked for both and should work for other robots and animals, too.

The biological robot, the roach, was the far superior runner with neurological signals guiding six impeccably evolved legs. The mechanical robot, a consumer model, had four stubby legs and no nervous system but relied instead for locomotion control on coarse physical forces traveling through its chassis as crude signals to roughly coordinate its clunky gait.

“The robot was much bulkier and could hardly sense its environment. The cockroach had many senses and can adapt better to rough terrain. Bumps as high as its hips wouldn’t slow it down at all,” said Izaak Neveln, the study’s first author, who was a postdoctoral researcher in the lab of Simon Sponberg at Georgia Techduring the study.

Advanced simplicity 

The method, or “measure,” as the study calls it, transcended these huge differences, which pervade animal-inspired robotics.

“The measure is general (universal) in the sense that it can be used regardless of whether the signals are neural spiking patterns, kinematics, voltages or forces and does not depend on the particular relationship between the signals,” the study’s authors wrote.

No matter how a bug or a bot functions, the measure’s mathematical inputs and outputs are always in the same units. The measure will not always eliminate the need for modeling, but it stands to shorten and guide modeling and avert anguishing missteps.

The authors published the study in the journal Nature Communications in August 2019. The research was funded by the National Science Foundation. Sponberg is an assistant professor in Georgia Tech’s School of Physics and in the School of Biological Sciences.

Centralization vs. decentralization

Often a bot or an animal sends many walking signals through a central system to harmonize locomotion, but not all signals are centralized. Even in humans, though locomotion strongly depends on signals from the central nervous system, some neural signals are confined to regions of the body; they are localized signals.

Some insects appear to move with little centralization -- such as stick bugs, also known as walking sticks, whose legs prod about nearly independently. Stick bugs are wonky runners.

“The idea has been that the stick bugs have the more localized control of motion, whereas a cockroach goes very fast and needs to maintain stability, and its motion control is probably more centralized, more clocklike,” Neveln said.

Strong centralization of signals generally coordinates locomotion better. Centralized signals could be code traveling through an elaborate robot’s wiring, a cockroach’s central neurons synching its legs, or the clunky robot's chassis tilting away from a leg thumping the ground thus putting weight onto an opposing leg.

Roboticists need to see through the differences and figure out the interplay of a locomotor’s local and central signals.

^{[Ready for graduate school? Here's how to apply to Georgia Tech.]} 

Cool physics

The new “measure” does this by focusing on an overarching phenomenon in the walking legs, which can be seen as pendula moving back and forth. For great locomotion, they need to synch up in what is called phase-coupling oscillations.

A fun, easy experiment illustrates this physics principle. If a few, say six, metronomes – ticking rhythm pendula that piano teachers use -- are swinging out of sync, and you place them all on a platform that freely sways along with the metronomes’ swings, the swings will sync up in unison.

The phases, or directions, of their oscillations are coupling with each other by centralizing their composite mechanical impulses through the platform. This particular example of phase-coupling is mechanical, but it can also be computational or neurological -- like in the roach.

Its legs would be analogous to the swinging metronomes, and central neuromuscular activity analogous to the free-swaying platform. In the roach, not all six legs swing in the same direction.

“Their synchronization is not uniform. Three legs are synchronized in phase with each other -- the front and back legs of one side with the middle leg of the other side -- and those three are synchronized out of phase with the other three,” Neveln said. “It’s an alternating tripod gait. One tripod of three legs alternates with the other tripod of three legs.”

Useless pogoing

And just like pendula, each leg’s swings can be graphed as a wave. All the legs’ waves can be averaged into an overall roach scurry wave and then developed into more useful math that relates centralization with decentralization and factors like entropy that can throw locomotion control off.

The resulting principles and math benefited the clunky robot, which has strong decentralized signals in its leg motors that react to leg contact with the ground, and centralized control weaker than that of the stick bug. The researchers graphed out the robot's movements, too, but they didn't result in the neatly synced group of waves that the cockroach had produced.

The researchers turned with the principles and math to the clunky robot, which initially was out of sorts -- bucking or hopping uselessly like a pogo stick. Then the scientists strengthened centralized control by re-weighting its chassis to make it move more coherently.

“The metronomes on the platform are mechanical coupling, and our robot coordinates control that way,” Neveln said. “You can change the mechanical coupling of the robot by repositioning its weights. We were able to predict the changes this would make by using the measure we developed from the cockroach.”

Cockroach surprises

The researchers also wired up specific roach muscles and neurons to observe their syncopations with the scurry waves. Seventeen cockroaches took 2,982 strides to inform the principles and math, and the bugs also sprung surprises on the researchers.

One stuck out: The scientists had thought signaling centralized more when the roach sped up, but instead, both central and local signaling strengthened, perhaps doubling down on the message: Run!

Georgia Tech’s Amoolya Tiramulai coauthored the paper. The National Science Foundation funded the research (grant # NSF CAREER MPS/PoLS 1554790). Any findings, conclusions, and recommendations are those of the authors and not necessarily of the NSF.

Writer & Media Representative: Ben Brumfield (404-660-1408), email: ben.brumfield@comm.gatech.edu

Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Flashback to July 2019, the dawn of autonomous vehicles and other connected cars, and physicists at the Georgia Institute of Technology and Multiscale Systems, Inc. have applied physics in a new study to simulate what it would take for future hackers to wreak exactly this widespread havoc by randomly stranding these cars. The researchers want to expand the current discussion on automotive cybersecurity, which mainly focuses on hacks that could crash one car or run over one pedestrian, to include potential mass mayhem.

They warn that even with increasingly tighter cyber defenses, the amount of data breached has soared in the past four years, but objects becoming hackable can convert the rising cyber threat into a potential physical menace.

“Unlike most of the data breaches we hear about, hacked cars have physical consequences,” said Peter Yunker, who co-led the study and is an assistant professor in Georgia Tech’s School of Physics.

It may not be that hard for state, terroristic, or mischievous actors to commandeer parts of the internet of things, including cars.

“With cars, one of the worrying things is that currently there is effectively one central computing system, and a lot runs through it. You don’t necessarily have separate systems to run your car and run your satellite radio. If you can get into one, you may be able to get into the other,” said Jesse Silverberg of Multiscale Systems, Inc., who co-led the study with Yunker 

Freezing traffic solid

In simulations of hacking internet-connected cars, the researchers froze traffic in Manhattan nearly solid, and it would not even take that to wreak havoc. Here are their results, and the numbers are conservative for reasons mentioned below.

“Randomly stalling 20 percent of cars during rush hour would mean total traffic freeze. At 20 percent, the city has been broken up into small islands, where you may be able to inch around a few blocks, but no one would be able to move across town,” said David Yanni, a graduate research assistant in Yunker’s lab.

Not all cars on the road would have to be connected, just enough for hackers to stall 20 percent of all cars on the road. For example, if 40 percent of all cars on the road were connected, hacking half would suffice.

Hacking 10 percent of all cars at rush hour would debilitate traffic enough to prevent emergency vehicles from expediently cutting through traffic that is inching along citywide. The same thing would happen with a 20 percent hack during intermediate daytime traffic.

The researchers’ results appear in the journal Physical Review E on July 20, 2019. The study is not embargoed.

^{[Ready for graduate school? Here's how to apply to Georgia Tech.]} 

It could take less

For the city to be safe, hacking damage would have to be below that. In other cities, things could be worse.

“Manhattan has a nice grid, and that makes traffic more efficient. Looking at cities without large grids like Atlanta, Boston, or Los Angeles, and we think hackers could do worse harm because a grid makes you more robust with redundancies to get to the same places down many different routes,” Yunker said.

The researchers left out factors that would likely worsen hacking damage, thus a real-world hack may require stalling even fewer cars to shut down Manhattan.

“I want to emphasize that we only considered static situations – if roads are blocked or not blocked. In many cases, blocked roads spill over traffic into other roads, which we also did not include. If we were to factor in these other things, the number of cars you’d have to stall would likely drop down significantly,” Yunker said.

The researchers also did not factor in ensuing public panic nor car occupants becoming pedestrians that would further block streets or cause accidents. Nor did they consider hacks that would target cars at locations that maximize trouble.

They also stress that they are not cybersecurity experts, nor are they saying anything about the likelihood of someone carrying out such a hack. They simply want to give security experts a calculable idea of the scale of a hack that would shut a city down.

The researchers do have some general ideas of how to reduce the potential damage.

“Split up the digital network influencing the cars to make it impossible to access too many cars through one network,” said lead author Skanka Vivek, a postdoctoral researcher in Yunker’s lab. “If you could also make sure that cars next to each other can’t be hacked at the same time that would decrease the risk of them blocking off traffic together.”

Traffic jams as physics

Yunker researches in soft matter physics, which looks at how constituent parts – in this case, connected cars – act as one whole physical phenomenon. The research team analyzed the movements of cars on streets with varying numbers of lanes, including how they get around stalled vehicles and found they could apply a physics approach to what they observed.

“Whether traffic is halted or not can be explained by classic percolation theory used in many different fields of physics and mathematics,” Yunker said.

Percolation theory is often used in materials science to determine if a desirable quality like a specific rigidity will spread throughout a material to make the final product uniformly stable. In this case, stalled cars spread to make formerly flowing streets rigid and stuck.

The shut streets would be only those in which hacked cars have cut off all lanes or in which they have become hindrances that other cars can’t maneuver around and do not include streets where hacked cars still allow traffic flow.

The researchers chose Manhattan for their simulations because a lot of data was available on that city’s traffic patterns.

Also READ: Georgia Tech's cybersecurity researchers tackle the internet of things 

The study was coauthored by Skanda Vivek and David Yanni of Georgia Tech and Jesse Silverberg of Multiscale Systems, Inc. Any findings, conclusions, and recommendations are those of the authors.

Writer & Media Representative: Ben Brumfield (404-660-1408), email: ben.brumfield@comm.gatech.edu

Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

The prototype robots respond to different vibration frequencies depending on their configurations, allowing researchers to control individual bots by adjusting the vibration. Approximately two millimeters long – about the size of the world’s smallest ant – the bots can cover four times their own length in a second despite the physical limitations of their small size.

“We are working to make the technology robust, and we have a lot of potential applications in mind,” said Azadeh Ansari, an assistant professor in the School of Electrical and Computer Engineering at the Georgia Institute of Technology. “We are working at the intersection of mechanics, electronics, biology and physics. It’s a very rich area and there’s a lot of room for multidisciplinary concepts.”

A paper describing the micro-bristle-bots has been accepted for publication in the Journal of Micromechanics and Microengineering. The research was supported by a seed grant from Georgia Tech’s Institute for Electronics and Nanotechnology. In addition to Ansari, the research team includes George W. Woodruff School of Mechanical Engineering Associate Professor Jun Ueda and graduate students DeaGyu Kim and Zhijian (Chris) Hao.

The micro-bristle-bots consist of a piezoelectric actuator glued onto a polymer body that is 3D-printed using two-photon polymerization lithography (TPP). The actuator generates vibration and is powered externally because no batteries are small enough to fit onto the bot. The vibrations can also come from a piezoelectric shaker beneath the surface on which the robots move, from an ultrasound/sonar source, or even from a tiny acoustic speaker.

The vibrations move the springy legs up and down, propelling the micro-bot forward. Each robot can be designed to respond to different vibration frequencies depending on leg size, diameter, design and overall geometry. The amplitude of the vibrations controls the speed at which the micro-bots move. 

“As the micro-bristle-bots move up and down, the vertical motion is translated into a directional movement by optimizing the design of the legs, which look like bristles,” explained Ansari. “The legs of the micro-robot are designed with specific angles that allow them to bend and move in one direction in resonant response to the vibration.”

The micro-bristle-bots are made in a 3D printer using the TPP process, a technique that polymerizes a monomer resin material. Once the portion of the resin block struck by the ultraviolet light has been chemically developed, the remainder can be washed away, leaving the desired robotic structure.

“It’s writing rather than traditional lithography,” Ansari explained. “You are left with the structure that you write with a laser on the resin material. The process now takes quite a while, so we are looking at ways to scale it up to make hundreds or thousands of micro-bots at a time.”

Some of the robots have four legs, while others have six. First author DeaGyu Kim made hundreds of the tiny structures to determine the ideal configuration.

The piezoelectric actuators, which use the material lead zirconate titanate (PZT), vibrate when electric voltage is applied to them. In reverse, they can also be used to generate a voltage, when they are vibrated, a capability the micro-bristle-bots could use to power up onboard sensors when they are actuated by external vibrations.

Ansari and her team are working to add steering capability to the robots by joining two slightly different micro-bristle-bots together. Because each of the joined micro-bots would respond to different vibration frequencies, the combination could be steered by varying the frequencies and amplitudes. “Once you have a fully steerable micro-robot, you can imagine doing a lot of interesting things,” she said.

Other researchers have worked on micro-robots that use magnetic fields to produce movement, Ansari noted. While that is useful for moving entire swarms at once, magnetic forces cannot easily be used to address individual robots within a swarm. The micro-bristle-bots created by Ansari and her team are believed to be the smallest robots powered by vibration.

The micro-bristle-bots are approximately two millimeters in length, 1.8 millimeters wide and 0.8 millimeters thick, and weigh about five milligrams. The 3D printer can produce smaller robots, but with a reduced mass, the adhesion forces between the tiny devices and a surface can get very large. Sometimes, the micro-bots cannot be separated from the tweezers used to pick them up.

Ansari and her team have built a “playground” in which multiple micro-bots can move around as the researchers learn more about what they can do. They are also interested in developing micro-bots that can jump and swim.

“We can look at the collective behavior of ants, for example, and apply what we learn from them to our little robots,” she added. “These micro-bristle-bots walk nicely in a laboratory environment, but there is a lot more we will have to do before they can go out into the outside world.”

The micro-bot fabrication was performed at the Georgia Tech Institute for Electronics and Nanotechnology, a member of the National Nanotechnology Coordinated Infrastructure, which is supported by the National Science Foundation through grant ECCS-1542173.

CITATION: DeaGyu Kim, Zhijian Hao, Jun Ueda and Azadeh Ansari, “A 5mg micro-bristle-bot fabricated by two-photon lithography” (Journal of Micromechanics and Microengineering, 2019). https://doi.org/10.1088/1361-6439/ab309b

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu).

Writer: John Toon

Hudgens holds a Ph.D. in ceramic engineering from Iowa State University. He has led research and development programs in national security, cybersecurity, quantum information science, and photonic microsystems. He also led programs in data analytics, synthetic aperture radar, and airborne intelligence, surveillance and reconnaissance (ISR) systems before becoming director of the $265 million-per-year TIC, which has a staff of 550 professionals working in six states and 136 different laboratories. 

A senior technology executive with 23 years of experience in national security research, Hudgens has also held positions at optical networking firm Mahi Networks, defense contractor Raytheon Electronic Systems, and semiconductor company Texas Instruments. In 2013, he won the Department of Energy Secretary’s Honor Award for Achievement for leading the Copperhead counter-IED program.

“Jim Hudgens has extensive experience building and leading federally sponsored programs that are at the center of GTRI’s core research areas,” said Chaouki Abdallah, Georgia Tech’s Executive Vice President for Research. “His experience developing and managing programs at Sandia National Laboratories and major private-sector defense contractors will support GTRI’s continued growth in service to our nation’s defense agencies and other important state and federal sponsors.”

GTRI has more than 2,300 employees conducting nearly $500 million worth of research across a broad range of technology areas that focus on solving critical challenges for government and industry sponsors. GTRI is one of the world’s leading applied research and development organizations, and is an integral part of Georgia Tech’s research program.

“Georgia Tech, through GTRI, is entrusted with a vital role in our national security,” Hudgens said. “I know firsthand that GTRI and other Georgia Tech researchers are known for the exceptional quality of their work in delivering innovative solutions to the most complex national security challenges.

“It is a great privilege for me to join the combined University System of Georgia and Georgia Tech family to develop a shared vision for how we will build on this reputation to advance one of the nation’s leading technological research universities,” he added. “I thank Georgia Tech President G.P. “Bud” Peterson, Provost Rafael Bras, and Executive Vice President Abdallah for the honor of becoming part of GTRI’s 85-year legacy of service to the state of Georgia and our nation.”

In congratulating Hudgens, Peterson emphasized GTRI’s important role in the nation, region, state – and Georgia Tech itself.

“For decades, the U.S. government and industry have looked to Georgia Tech – in particular GTRI – as they seek to find and develop effective, creative solutions in national security and other mission-critical areas,” Peterson said. “We are pleased to welcome Jim Hudgens to lead one of Georgia Tech’s most important missions in support of our nation, region, and state.”

Hudgens’ selection came after a five-month national search during which he was one of four finalists to make presentations to Georgia Tech faculty and staff.

Sandia National Laboratories is a multi-mission laboratory operated for the U.S. Department of Energy’s National Nuclear Security Administration. Sandia has major research and development responsibilities in nuclear deterrence, global security, defense, energy technologies, and economic competitiveness, with main facilities in Albuquerque, New Mexico, and Livermore, California. Sandia is the largest of the country’s 17 national laboratories.

GTRI conducts research through eight laboratories located on Georgia Tech’s midtown Atlanta campus, in a research facility near Dobbins Air Reserve Base in Smyrna, Georgia, and in Huntsville, Alabama. GTRI also has more than a dozen locations around the nation where it serves the needs of its research sponsors. GTRI’s research spans a variety of disciplines, including autonomous systems, cybersecurity, electromagnetics, electronic warfare, modeling and simulation, sensors, systems engineering, test and evaluation, and threat systems.

Media Relations Assistance: John Toon (404-894-6986) (jtoon@gatech.edu).

Writer: John Toon

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Powered by a pair of photovoltaic panels and designed to linger in the forest canopy continuously for months, SlothBot moves only when it must to measure environmental changes – such as weather and chemical factors in the environment – that can be observed only with a long-term presence. The proof-of-concept hyper-efficient robot, described May 21 at the International Conference on Robotics and Automation (ICRA) in Montreal, may soon be hanging out among treetop cables in the Atlanta Botanical Garden.

“In robotics, it seems we are always pushing for faster, more agile and more extreme robots,” said Magnus Egerstedt, the Steve W. Chaddick School Chair of the School of Electrical and Computer Engineering at the Georgia Institute of Technology and principal investigator for Slothbot. “But there are many applications where there is no need to be fast. You just have to be out there persistently over long periods of time, observing what’s going on.”

Based on what Egerstedt called the “theory of slowness,” Graduate Research Assistant Gennaro Notomista designed SlothBot together with his colleague, Yousef Emam, using 3D-printed parts for the gearing and wire-switching mechanisms needed to crawl through a network of wires in the trees. The greatest challenge for a wire-crawling robot is switching from one cable to another without falling, Notomista said.

“The challenge is smoothly holding onto one wire while grabbing another,” he said. “It’s a tricky maneuver and you have to do it right to provide a fail-safe transition. Making sure the switches work well over long periods of time is really the biggest challenge.”

Mechanically, SlothBot consists of two bodies connected by an actuated hinge. Each body houses a driving motor connected to a rim on which a tire is mounted. The use of wheels for locomotion is simple, energy efficient and safer than other types of wire-based locomotion, the researchers say.

SlothBot has so far operated in a network of cables on the Georgia Tech campus. Next, a new 3D-printed shell – that makes the robot look more like a sloth – will protect the motors, gears, actuators, cameras, computer and other components from the rain and wind. That will set the stage for longer-term studies in the tree canopy at the Atlanta Botanical Garden, where Egerstedt hopes visitors will see a SlothBot monitoring conditions as early as this fall.

The name SlothBot is not a coincidence. Real-life sloths are small mammals that live in jungle canopies of South and Central America. Making their living by eating tree leaves, the animals can survive on the daily caloric equivalent of a small potato. With their slow metabolism, sloths rest as much 22 hours a day and seldom descend from the trees where they can spend their entire lives.

“The life of a sloth is pretty slow-moving and there’s not a lot of excitement on a day-to-day level,” said Jonathan Pauli, an associate professor in the Department of Forest & Wildlife Ecology at the University of Wisconsin-Madison, who has consulted with the Georgia Tech team on the project. “The nice thing about a very slow life history is that you don’t really need a lot of energy input. You can have a long duration and persistence in a limited area with very little energy inputs over a long period of time.”

That’s exactly what the researchers expect from SlothBot, whose development has been funded by the U.S. Office of Naval Research.

“There is a lot we don’t know about what actually happens under dense tree-covered areas,” Egerstedt said. “Most of the time SlothBot will be just hanging out there, and every now and then it will move into a sunny spot to recharge the battery.”

The researchers also hope to test SlothBot in a cacao plantation in Costa Rica that is already home to real sloths. “The cables used to move cacao have become a sloth superhighway because the animals find them useful to move around,” Egerstedt said. “If all goes well, we will deploy SlothBots along the cables to monitor the sloths.”

Egerstedt is known for algorithms that drive swarms of small wheeled or flying robots. But during a visit to Costa Rica, he became interested in sloths and began developing what he calls “a theory of slowness” together with Professor Ron Arkin in Georgia Tech’s School of Interactive Computing. The theory leverages the benefits of energy efficiency.

“If you are doing things like environmental monitoring, you want to be out in the forest for months,” Egerstedt said. “That changes the way you think about control systems at a high level.”

Flying robots are already used for environmental monitoring, but their high energy needs mean they cannot linger for long. Wheeled robots can get by with less energy, but they can get stuck in mud or be hampered by tree roots, and cannot get a big picture view from the ground.

“The thing that costs energy more than anything else is movement,” Egerstedt said. “Moving is much more expensive than sensing or thinking. For environmental robots, you should only move when you absolutely have to. We had to think about what that would be like.”

For Pauli, who studies a variety of wildlife, working with Egerstedt to help SlothBot come to life has been gratifying.

“It is great to see a robot inspired by the biology of sloths,” he said. “It has been fun to share how sloths and other organisms that live in these ecosystems for long periods of time live their lives. It will be interesting to see robots mirroring what we see in natural ecological communities.”

This research was sponsored by the U.S. Office of Naval Research through Grant N00014-15-2115. The content is solely the responsibility of the authors and does not necessarily represent the official views of the ONR.

CITATION: "The SlothBot: A Novel Design for a Wire-Traversing Robot," IEEE Robotics and Automation Letters, (Volume 4, Issue 2, April 2019) https://ieeexplore.ieee.org/document/8642808

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181  USA

Media Relations Contact: John Toon (404-894-6986) (jtoon@gatech.edu)

Writer: John Toon

DoD’s Multidisciplinary University Research Initiatives (MURI) Program funds projects that bring researchers together from diverse backgrounds to work on a complex problem. Institute for Data Engineering and Science co-director, Professor Dana Randall, is project investigator and leads a team of six that includes Daniel Goldman, Dunn Family Professor in the School of Physics. The Formal Foundations of Algorithmic Matter and Emergent Computation team also includes chemical engineering, mechanical engineering, physics, and computational science researchers from other universities.  

The researchers are trying to predict and design emergent behavior within computation by using basic algorithms on simple machines to perform complex tasks. Emergent behavior is when a microscopic change in a parameter creates a macroscopic change to a system. This collective behavior is easy to find in nature, from a swarm of bees to a colony of ants, but also appears in other scientific disciplines. 

“A MURI lets us take a deep dive toward understanding how many computationally limited components at the micro-scale can be programmed to work collectively to produce useful behavior at the macro-scale,” said Randall, who is also the ADVANCE Professor of Computing. “Our interdisciplinary team combines expertise in many fields, mimicking the research by forming a collaboration that is also greater than the sum of its parts."

The MURI hybrid approach to algorithmic matter combines traditional logic-based programming with non-traditional computational methods, such as using physical characteristics of the interacting matter to drive a system toward collective behavior. One of the goals is to program based on this predictable emergent behavior. The approach also predicts basic properties of the collective’s emergent behavior, like whether it will behave like a gas, fluid, or solid. In this context, emergent behavior turns into emergent collective computation.

“MURI promises basic algorithms that allow very simple machines to work collectively to perform amazingly complex tasks,” Massachusetts Institute of Technology (MIT) chemical engineering Professor Michael Strano said. “Our team will examine systems of autonomous cell-like particles that interact and respond to the movement of their neighbors in a programmable way. Theorists will be able to test ideas of emergent computation from these simple devices and learn how to execute tasks from the behavior of relatively simple, autonomous particles.”

Although the behavior has footing in physics, computer science, and swarm robotics, there is no underlying framework to explain why until this research. The multidisciplinary approach allows theory and experiment to continuously inform each other and determine the computational capabilities of emergent behavior. The team has an ideal range of expertise in machine learning, control theory, and non-equilibrium physics and algorithms. They are also working with experimentalists who build collective systems at granular and microscopic scales.  

“An exciting aspect of this collaboration will be our attempts to interface and integrate ideas and tools from robotics, non-equilibrium physics, control theory, and computer science to develop task-capable swarms,” Goldman said.

This MURI project will run for five years and is funded by the Army Research Office. In addition to Randall, Goldman, and Strano, the team also includes Arizona State computational science and engineering Professor Andrea Richa, MIT physics Associate Professor Jeremy England, and Northwestern mechanical engineering Professor Todd Murphey.

The overarching goal is to find how simplistic the computation can be for this complexity. This could lead to advances in engineered systems achieving specific task-oriented goals.

“The MURI promises nothing short of the transformation of robots,” Strano said, “from the large, bulky constructions that we think of today, to future clouds or swarms that enable functions that are currently impossible to realize.”

Writer: Tess Malone