Coordinating Electrical Stimulation and Motor Assist in a Hybrid Neuroprosthesis Using Control Strategies Inspired by Human Motor Control

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

Functional electrical stimulation (FES) and powered exoskeletons are two technologies being used to restore walking in individuals with paraplegia. FES comprises low-level electrical currents applied to activate leg muscles. In contrast, powered exoskeletons use electric motors mounted on an external wearable frame to move lower-limb joints. Each of these technologies has limitations. In this project they are used in synergy, to create a hybrid neuroprosthesis that addresses these individual drawbacks. Critical to the success of the hybrid approach is coordinated control of multiple FES-activated muscles and the electric motors. This coordinated control must adapt over time as FES-induced fatigue degrades the ability of the user's muscles to follow the desired walking motion. The control algorithms resulting from this project will enable consistent walking movements despite FES-induced muscle fatigue, contributing to the emergence of an adaptable and lightweight hybrid exoskeleton with substantial advantages over FES systems or powered exoskeletons alone. This research will enhance physical activity and improve mobility for individuals with impaired lower limb function, enabling greater community participation and increased quality of life. In a hybrid exoskeleton, coordinating multiple FES-activated muscles and electric motors can be complicated due to redundancy. Further, updating multiple control inputs to account for system uncertainty and FES-induced muscle fatigue can be computationally expensive for the real-time control. This research will design adaptive low-dimensional control methods for the hybrid neuroprosthesis. The human motor control inspired control structure can be used to control multiple effector system using a fewer number of commands. Thus, redundancy and complexity associated with the closed-loop control of the hybrid neuroprosthesis will be reduced. Dynamic optimization will be used to design low-dimensional control modules for walking. Then by using Lyapunov-based stability analysis, update and feedback control laws for the control structure will be designed. This will ensure stability and tracking despite system uncertainty and muscle fatigue. Finally, the new control structure will be experimentally verified on able-bodied subjects.