Echo state network based predictive control with particle swarm optimization for pneumatic muscle actuator

To realize a high-accurate trajectory tracking control of the Pneumatic Muscle Actuator (PMA), a comprehensive single-layer neural network (SNN) and Echo State Neural Network (ESN) based predictive control with particle swarm optimization (PSO) is proposed, where PSO optimizes the weight coefficients of the SNN and the ESN state is updated by the online Recursive Least Square (RLS) algorithm for predictive control. Based on Lyapunov theory, the learning convergence theorem is established to guarantee the stability of the closed-loop system. The proposed control algorithm is employed for the trajectory tracking control of PMA. The interface between the xPC target and the virtual instrument was established to realize the real-time control and to make the control more accurate and stable. Both simulations and experiments were performed to verify the proposed methods. The experiments were conducted on the real PMA system, which was connected with the xPC target system. The results demonstrate the validity of PMA as well as the effectiveness of the novel control algorithm.