Fuzzy-Sliding Mode Control for Humanoid Arm Robots Actuated by Pneumatic Artificial Muscles With Unidirectional Inputs, Saturations, and Dead Zones

Recently, the pneumatic artificial muscle (PAM) that can reproduce natural muscle functionalities has become one of the core actuator mechanisms of intelligent interactive soft robots. Unfortunately, some inherent defects (e.g., unignorable nonlinearities, hysteresis, low shrinkage frequencies, etc.) have limited the application progress of humanoid PAM arm robots. Additionally, the input constraints (e.g., saturations, dead zones, unidirectional inputs, etc.), unexpected external disturbances, unidentifiable system parameters, and inevitable unmodeled dynamics are usually complicated, which cannot be easily eliminated through existing adaptive control methods. This article proposes an adaptive fuzzy-sliding mode control method for humanoid PAM arm robots without any information of precise model structures and system parameters, which can suppress the unexpected effects of complicated unknown functions and achieve high performance tracking control, simultaneously. To the best of our knowledge, the proposed controller is the first method for the humanoid PAM arm robots that considers the nonlinear input constraints including unidirectional conditions, saturations, and dead zones, simultaneously. Next, all the input constraints, system parameter uncertainties, unmodeled dynamics, and external disturbances can be estimated adaptively by utilizing the proposed fuzzy update law. Particularly, a sliding mode control law is designed to compensate possible fuzzy approximation errors, and rigorous Lyapunov-based stability analysis is provided to ensure that the state errors can converge to zero within finite time. Hardware experiments are carried out later tovalidate the effectiveness and robustness of the proposed method.