Title: Sensor Fusion Representation of Locomotion Biomechanics with Applications in the Control of Lower Limb Prosthesis

Date: Wednesday, July 7, 2021

Time: 12:00 pm - 2:00 pm (EST)

Location: GTMI Auditorium

Virtual access: BlueJeans

Link: https://bluejeans.com/209922501/9690

Meeting ID: 209 922 501

Participant Passcode: 9690

Jonathan Camargo

Robotics Ph.D. Candidate

School of Mechanical Engineering

Georgia Institute of Technology

Committee:

Dr. Aaron Young (Advisor), School of Mechanical Engineering, Georgia Institute of Technology

Dr. Omer Inan, School of Electrical and Computer Engineering, Georgia Institute of Technology

Dr. Boris Prilutsky, School of Biological Sciences, Georgia Institute of Technology

Dr. Gregory Sawicki, School of Mechanical Engineering, Georgia Institute of Technology

Dr. Ye Zhao, School of Mechanical Engineering, Georgia Institute of Technology

 Abstract:

Free locomotion and movement in diverse environments are significant concerns for individuals with amputation who need independence in daily living activities. As users perform community ambulation, they face changing contexts that challenge what the typical passive prosthesis can offer. This problem rises opportunities for developing intelligent robotic systems that assist the locomotion with the least possible interruptions for direct input during operation.

The use of multiple sensors to detect the user's intent and locomotion parameters is a promising technique that could provide a fast and natural response to the prostheses. However, the use of these sensors still requires a thorough investigation before they can be translated into practical settings. In addition, the dynamic change of context during locomotion should translate to adjustment in the device's response. To achieve the scaling rules for this modulation, a rich biomechanics dataset of community ambulation would provide a source of quantitative criteria to generate bioinspired controllers.

This dissertation produces a better understanding of the characteristics of community ambulation from two different perspectives: the biomechanics of human motion and the sensory signals that can be captured by wearable technology.

By studying human locomotion in diverse environments, including walking on stairs, ramps, and level ground, this work generated a comprehensive open-source dataset containing the biomechanics and signals from wearable sensors during locomotion, evaluating the effects of changing the locomotion context within the ambulation mode. With the multimodal dataset, I developed and evaluated a combined strategy for ambulation mode classification and the estimation of locomotion parameters, including the walking speed, stair height, ramp slope, and biological moment. Finally, by combining this knowledge and incorporating both the biomechanics insight with the machine learning-based inference in the frame of impedance control, I propose novel methods to improve the performance of lower-limb robotics with a focus on powered prostheses.