Functional electrical stimulation of a quadriceps muscle using a neural-network adaptive control approach

Functional electrical stimulation (FES) has been used to facilitate persons with paralysis in restoring their motor functions. In particular, FES-based devices apply electrical current pulses to stimulate the intact peripheral nerves to produce artificial contraction of paralyzed muscles. The aim of this work is to develop a model reference adaptive controller of the shank movement via FES. A mathematical model, which describes the relationship between the stimulation pulsewidth and the active joint torque produced by the stimulated muscles in non-isometric conditions, is adopted. The direct adaptive control strategy is used to address those nonlinearities which are linearly parameterized (LP). Since the torque due to the joint stiffness component is non-LP, a neural network (NN) is applied to approximate it. A backstepping approach is developed to guarantee the stability of the closed loop system. In order to address the saturation of the control input, a model reference adaptive control approach is used to provide good tracking performance without jeopardizing the closed-loop stability. Simulation results are provided to validate the proposed work.