A Method for FES Control of Human Knee Joint with Time-dependent Model Parameters

Functional electrical stimulation (FES) is one useful neuroprosthetic technique to restore motor function for motion-impaired patients. Muscle contractions can be artificially driven through delivering electrical short pulses to targeted motor neurons and muscles. FES has been frequently applied in human movement control to fulfill various rehabilitation training tasks, and many open-loop and closed-loop paradigms based on human neuromuscular-skeletal models have been established. However, one common issue which appears in FES control of movement is that the human plant to be controlled is with time-dependent (time-varying) model parameters. Such time-vary factor may often vitiate accurate FES control for human joint movement, and the tracking control performance seems unsatisfactory via conventional control approaches which are originally designed for models with constant (static) coefficients. To overcome the shortcomings of current conventional FES control methods and seek for new control strategies of dealing with time-varying issues in FES, this paper aims to propose a novel approach for FES control of human knee joint with time-varying model parameters. The parameters which appear in the joint dynamics are analytically assumed to be or estimated as time-varying rather than static. Simulation results verify the promising performance by the proposed FES control method for such time-varying knee joint dynamics model.