Control of intraspinal microstimulation using an adaptive terminal-based neuro-sliding mode control

In this study, a terminal-based neuro-sliding mode control was presented for control of ankle joint movement using intraspinal microstimulation (ISMS). The controller was based on the combination of an adaptive terminal-based neural controller with a terminal-based sliding mode control (SMC). To implement the SMC, the dynamics of plant to be controlled should be presented in a canonical form. In this work, a terminal sliding mode based estimation method was proposed for estimation of the unknown dynamics of plant. The most advantage of terminal sliding mode is the finite-time convergence compared to the conventional gradient descent (GD) learning algorithm. The experiments were conducted on two Wistar rats. Microelectrodes were inserted into the motor pools of dorsiflexor and plantar flexor muscles. The results show that good tracking performance was obtained using this control scheme for three periods of gait cycle. The average of tracking error over all trials of experiments and all movement periods for two rats is 6.8° (9.8%).