{"690590":{"#nid":"690590","#data":{"type":"news","title":"Engineering Safer Skies","body":[{"value":"\u003Cp\u003EAs the skies grow more congested, from low Earth orbit to downtown airspace, the need for smarter autonomous systems has never been greater. Collisions in space have become an expensive problem, and urban air mobility is an option worth exploring given the increasing traffic problems plaguing major cities.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EManaging these complexities is a mathematical challenge and the \u003Ca href=\u0022https:\/\/sites.gatech.edu\/c3uae\/\u0022\u003E\u003Cstrong\u003EControl, Coordination, and Competition under Uncertainty (C3U) Lab\u003C\/strong\u003E\u003C\/a\u003E is tackling it head-on. Led by Assistant Professor \u003Ca href=\u0022https:\/\/ae.gatech.edu\/directory\/person\/sarah-hq-li\u0022\u003E\u003Cstrong\u003ESarah H.Q. Li\u003C\/strong\u003E\u003C\/a\u003E, the lab develops algorithms to help satellites make decisions to avoid orbital conflicts. She\u2018s developed an algorithm that allows a servicing satellite to assist other satellites that find themselves spinning out of control. Li is exploring airspace management for urban mobility, as well. She specializes in multi-agent systems and in interactive decision-making among autonomous vehicles, humans, and aircraft.\u0026nbsp;\u003C\/p\u003E\u003Cdiv\u003E\u003Cdiv\u003E\u003Ch3\u003E\u003Cstrong\u003EPreventing Debris from Derailing the Mission\u003C\/strong\u003E\u003C\/h3\u003E\u003Cp\u003EColliding with debris can end a satellite\u2019s mission in an instant, causing significant damage or sending the spacecraft into an uncontrollable spin. With growing congestion in low Earth orbit, other objects in space can put satellites at risk, and a poorly timed maneuver, to avoid space debris or another spacecraft, can be just as dangerous as not maneuvering at all. Li\u2019s research helps autonomous satellites reason through these encounters, coordinate with servicing satellites when needed, and decide when it\u2019s safest to move or hold position.\u003C\/p\u003E\u003Cp\u003E\u201cSatellite operators haven\u2019t established methods to automate collision\u2011avoidance maneuvers,\u201d Li said. \u201cA major challenge is the significant uncertainty that space operators must manage. Uncertainty in space exceeds what ground\u2011based autonomous systems encounter.\u201d\u003C\/p\u003E\u003Cp\u003EOn Earth, autonomous vehicles generally know their precise location, so the problem becomes determining how far or in which direction to move. In space, position estimates can span hundreds of kilometers. Often, operators on the ground simply cannot determine whether two objects will collide. To move or not to move becomes a critical question, especially with large differences in fuel costs. \u0026nbsp;Frequent moves can quickly deplete a satellite\u2019s limited fuel. Li\u2019s algorithm incorporates guidance used for ground\u2011based autonomous systems that\u2019s modified to account for these challenges. Li\u2019s\u0026nbsp;\u003Ca href=\u0022https:\/\/arxiv.org\/abs\/2508.05876\u0022\u003E\u003Cstrong\u003Epaper\u003C\/strong\u003E\u003C\/a\u003E presented at the\u0026nbsp;American Institute of Aeronautics and Astronautics (AIAA) Astrodynamics Specialist 2025 Conference detailing this work.\u003C\/p\u003E\u003Cp\u003EBeyond collision avoidance, satellites must also survive an impact. When a satellite is hit by debris or accumulates momentum over time, it can begin spinning unpredictably. Once that happens, stabilizing the satellite becomes extremely complicated.\u003C\/p\u003E\u003Cp\u003EIn a recent AIAA SciTech Forum \u003Ca href=\u0022https:\/\/arc.aiaa.org\/doi\/abs\/10.2514\/6.2026-0407\u0022\u003E\u003Cstrong\u003Epaper\u003C\/strong\u003E\u003C\/a\u003E\u003Cstrong\u003E,\u0026nbsp;\u003C\/strong\u003ELi and her team demonstrated a momentum\u2011balanced contact strategy that enables a servicing satellite to safely approach and stabilize a free\u2011spinning object. They simulated a it equipped with a robotic arm that approaches the spinning object and matches its rotation. Once the two are synchronized, the robotic arm extends and grabs the spinning object and stabilizes its orientation.\u003C\/p\u003E\u003Cp\u003EThey\u2019re approaching this research in multiple stages. The core algorithm is already functional, so the team is building a higher\u2011fidelity simulation environment to validate their approach. Once they validate the algorithm through additional experiments, they will transition to physical tests using drones.\u0026nbsp;\u003C\/p\u003E\u003C\/div\u003E\u003C\/div\u003E\u003Cdiv\u003E\u003Cdiv\u003E\u003Ch3\u003E\u003Cstrong\u003ENavigating City Skyways\u003C\/strong\u003E\u003C\/h3\u003E\u003Cp\u003EIn addition, C3U is focused on urban mobility issues with shared air space. As traffic negatively impacts major cities, and supply and demand for goods increases, the airways become a major consideration for transportation in urban areas. Once more urban airways are open, much like highways, there must be a system to avoid collision and confusion.\u003Cbr\u003E\u0026nbsp;In shared airspace, multiple drones, each traveling to their own destination, must navigate potential conflicts as they cross one another\u2019s paths. To manage this, Li applies principles from auction theory. In this framework, drones \u201cbid\u201d for access to specific portions of the airspace based on vehicle-specific preferences like specific altitudes, routes and visibility.\u003C\/p\u003E\u003Cp\u003E\u201cAn auctioneer\u2011like mechanism, similar to an air traffic controller, evaluates these bids and assigns routes accordingly, sometimes granting a preferred path, and other times determining that no safe route is available. The system is designed to encourage each drone to bid truthfully according to its actual needs, preventing both overbidding and underbidding,\u201d Li explained.\u003C\/p\u003E\u003Cp\u003EBy ensuring truthful participation, the mechanism can allocate airspace efficiently and fairly, allowing individually optimal decisions to produce coordinated, system\u2011wide safety.\u003C\/p\u003E\u003Cp\u003E\u201cEconomic models typically assume a direct link between price and reward, but engineering doesn\u2019t work that way,\u201d Li said. \u201cIn our world, everything depends on dynamics\u2014how vehicles move, how routes connect, and how safety constraints interact in real time. Our inputs are acceleration vectors to the drone, which controls how the drone moves through space, and the outputs are where it ends up. Only after that can we determine the true \u2018reward.\u2019 Engineering forces us to account for complex physical behavior. Part of my work is figuring out how to bridge those two worlds.\u201d\u003C\/p\u003E\u003Ch3\u003E\u003Cstrong\u003EAI Anticipating Human Intent\u003C\/strong\u003E\u003C\/h3\u003E\u003Cp\u003EC3U is also exploring how large language models (LLMs), advanced systems designed to understand and generate human language, can help autonomous aircraft understand human pilots. They are developing an LLM\u2011enabled system that can listen to real\u2011time radio communications and interpret pilot intentions, allowing autonomous aircraft to anticipate pilot in unpredictable environments rather than simply reacting to it.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EIn a recent \u003Ca href=\u0022https:\/\/arxiv.org\/abs\/2509.14063\u0022\u003E\u003Cstrong\u003Epaper\u003C\/strong\u003E\u003C\/a\u003E Li co-authored, she examined how LLMs can predict and improve landings at airports without a control tower.\u0026nbsp;\u003Cbr\u003E\u003Cbr\u003E\u201cWe are looking at using LLMs to process information so autonomous systems can infer what pilots are planning,\u201d Li explained. \u201cBy converting unstructured audio into structured content, the system allows autonomous aircraft to plan more safely.\u201d\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EFor Li, the project represents another step toward a future where autonomous vehicles can navigate complex, uncertain environments by understanding dynamics and the human language shaping the airspace around them.\u003C\/p\u003E\u003C\/div\u003E\u003C\/div\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003E\u003Ca\u003EThe C\u00b3U Lab\u0026nbsp;\u003C\/a\u003E develops algorithms that allow satellites and planes to make complex decisions in uncertain environments.\u0026nbsp;\u003Cbr\u003E\u003Cbr\u003E\u0026nbsp;\u003C\/p\u003E\u003Cdiv\u003E\u003Cdiv\u003E\u003Cdiv\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003C\/div\u003E\u003C\/div\u003E\u003C\/div\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Sarah Li creates technology that helps satellites and aircraft make smart, real\u2011time decisions in crowded, unpredictable conditions."}],"uid":"36345","created_gmt":"2026-06-02 13:10:25","changed_gmt":"2026-06-02 13:22:01","author":"gwaddell3","boilerplate_text":"","field_publication":"","field_article_url":"","location":"Atlanta, GA","dateline":{"date":"2026-06-02T00:00:00-04:00","iso_date":"2026-06-02T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"680400":{"id":"680400","type":"image","title":"Sarah-li-lab.jpg","body":"\u003Cp\u003E\u003Ca href=\u0022https:\/\/sites.gatech.edu\/c3uae\/\u0022\u003EControl, Coordination, and Competition under Uncertainty (C3U) Lab\u003C\/a\u003E\u0026nbsp;\u003C\/p\u003E","created":"1780405947","gmt_created":"2026-06-02 13:12:27","changed":"1780405947","gmt_changed":"2026-06-02 13:12:27","alt":"c3u lab","file":{"fid":"264656","name":"Sarah-li-lab.jpg","image_path":"\/sites\/default\/files\/2026\/06\/02\/Sarah-li-lab.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/06\/02\/Sarah-li-lab.jpg","mime":"image\/jpeg","size":223705,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/06\/02\/Sarah-li-lab.jpg?itok=RlF2rDjA"}},"680399":{"id":"680399","type":"image","title":"chart-cropped.png","body":"\u003Cp\u003EPredictions from GoalPredictor and TrajAirNet vs. true path of N624AQ\u003C\/p\u003E","created":"1780405841","gmt_created":"2026-06-02 13:10:41","changed":"1780405841","gmt_changed":"2026-06-02 13:10:41","alt":"Predictions from GoalPredictor and TrajAirNet vs. true path of N624AQ","file":{"fid":"264655","name":"chart-cropped.png","image_path":"\/sites\/default\/files\/2026\/06\/02\/chart-cropped.png","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/06\/02\/chart-cropped.png","mime":"image\/png","size":383507,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/06\/02\/chart-cropped.png?itok=jwd1XgoI"}}},"media_ids":["680400","680399"],"related_links":[{"url":"https:\/\/ae.gatech.edu\/news\/2025\/03\/georgia-tech-collaborate-67-million-nasa-university-leadership-initiative","title":"Georgia Tech to Collaborate on $6.7 Million NASA University Leadership Initiative"},{"url":"https:\/\/ae.gatech.edu\/news\/2024\/10\/ae-professors-research-aims-improve-decision-making-artificial-intelligence","title":"AE Professor\u2019s Research Aims to Improve Decision-Making in Artificial Intelligence"}],"groups":[{"id":"1239","name":"School of Aerospace Engineering"}],"categories":[{"id":"136","name":"Aerospace"}],"keywords":[{"id":"1325","name":"aerospace"},{"id":"179801","name":"urban air mobility"},{"id":"169609","name":"satellite"},{"id":"189101","name":"algorithm design"}],"core_research_areas":[{"id":"193655","name":"Artificial Intelligence at Georgia Tech"}],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EMonique Waddell\u003C\/p\u003E","format":"limited_html"}],"email":["monique.waddell@gatech.edu"],"slides":[],"orientation":[],"userdata":""}}}