RLS-ESN based PID control for rehabilitation robotic arms driven by PM-TS actuators

To drive a single joint of rehabilitation robotic arm, we propose a new PM-TS actuator comprising a Pneumatic Muscle (PM) and a Torsion Spring (TS). Unlike the traditional agonist/antagonist PM actuator, the PM is arranged in appropriate place as agonist and the torsion spring provides opposing torque as antagonist in the proposed actuator. The 1-DOF and 2-DOF rehabilitation robotic arm models are derived considering the PM-TS dynamic model. To realize a high-accurate trajectory tracking control of the robotic arms, an intelligent PID controller based on an Echo State Neural Network (ESN) is proposed, where the ESN state is updated by the online Recursive Least Square (RLS) algorithm. Simulation results demonstrate the validity of PM-TS actuators. The performance of RLS-ESN based PID controller is found more satisfactory than conventional PID controller in our study.