An Integrated Intelligent Nonlinear Control Method for a Pneumatic Artificial Muscle

Abstract

This paper proposes an advanced position-tracking control approach, referred to as an integrated intelligent nonlinear controller, for a pneumatic artificial muscle (PAM) system. Due to the existence of uncertain, unknown, and nonlinear terms in the system dynamics, it is difficult to derive an exact mathematical model with robust control performance. To overcome this problem, the main contributions of this paper are as follows. To actively represent the behavior of the PAM system using a grey-box model, neural networks are employed as equivalent internal dynamics of the system model and optimized online by a Lyapunov-based method. To realize the control objective by effectively compensating for the estimation error, an advanced robust controller is developed from the integration of the designed networks, and improvement of the sliding mode and backstepping techniques. The convergences of both the developed model and the closed-loop control system are guaranteed by Lyapunov functions. As a result, the overall control approach is capable of ensuring the system's performance with fast response, high accuracy, and robustness. Real-time experiments are carried out in a PAM system under different conditions to validate the effectiveness of the proposed method.