Title: A Framework for Autonomous Exoskeleton Assistance Independent of Activity

Date: Monday, December 13, 2021

Time: 10AM EST

Location: GTMI 201

Dean Molinaro

Robotics PhD Student

School of Mechanical Engineering

Georgia Institute of Technology

Committee:

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

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

Dr. Matthew Gombolay – School of Interactive Computing, Georgia Institute of Technology

Dr. Sehoon Ha – School of Interactive Computing, Georgia Institute of Technology

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

Abstract:

Robotic lower-limb exoskeletons have outstanding potential to improve human mobility, leading to increases in independence and quality of life. To date, hip exoskeletons have enhanced user mobility during controlled in-lab settings; however, the lack of an exoskeleton controller able to autonomously adapt with changes in user, activity, and activity intensity limits the viability of these systems in daily life. In this proposal, a novel hip exoskeleton controller is introduced, which commands assistance based on instantaneous estimates of the user’s hip flexion/extension moment via a temporal convolutional network. By modulating exoskeleton assistance based on hip moments, the controller uses a single, continuous variable to account for changes in user, activity, and intensity, removing the limitations of previous exoskeleton controllers. To maximize the performance and generalizability of the hip moment estimator, the model structure (Aim 1) and training set activities (Aim 3) will be optimized. To study human response to this framework, user metabolic cost during multimodal ambulation (Aim 2) and lower-limb muscle effort during an expanded activity set, including lunging and squatting, (Aim 3) will be quantified. Thus, this proposal introduces a first-of-its-kind exoskeleton controller that autonomously customizes assistance regardless of activity, bridging the gap between in-lab and real-world settings.