Brendon Allen has some exoskeletons in his closet, and the National Science Foundation (NSF) wants to learn more about them.

The NSF recently awarded the assistant professor in the Department of Mechanical Engineering a five-year $588,408 NSF CAREER Award.

Allen aims to increase access to rehabilitation for individuals with movement disorders through a deep learning control framework for home-based hybrid exoskeletons.

“These devices combine functional electrical stimulation (FES) with actuated robots to provide personalized therapy,” Allen said. “Shifting the computational demand from individual homes to clinicians’ offices can, I believe, reduce the cost of telerehabilitation significantly.”

According to Allen, communication delays between the clinician’s computer and the exoskeleton can exacerbate the device’s inherently uncertain and nonlinear dynamics.

“Those delays,” Allen said, “equal destabilization.”

Allen feels that destabilization can be drastically decreased utilizing novel delay compensation and deep neural network–based methods to enable the remote control of the exoskeleton.

“Successful completion of this project could transform the rehabilitation industry,” Allen said. “Not only will it make rehabilitation more accessible and affordable, but the control developments in this work will also improve centralized network control systems, impacting fields such as manufacturing, power grid automation, reconnaissance and search and rescue operations.”