Active Modeling and Control of the Ring-Shaped Pneumatic Actuator: An Experimental Study

Abstract

A soft ring-shaped pneumatic actuator (RPA) that contracts radially plays a key role in biomimicking the contractions of the digestive system. However, the modeling and control of the RPA are still challenging due to the nonlinearity, hysteresis, and unique structure. In this article, we present an active-model-enhanced control (AMEC) scheme for the precise contraction control of the RPA. First, a generalized mass–spring–damper model is proposed to describe the dynamics of the RPA focusing on the midpoint of the inner wall. Based on the proposed model, the AMEC scheme is developed to enhance a proportional–integral–derivative (PID) controller in terms of tracking a reference trajectory. In this scheme, the modeling errors of the generalized mass–spring–damper model are estimated by an active modeling algorithm, which is integrated into the PID controller to eliminate the effect of the modeling errors. Finally, experiments are conducted to validate the proposed modeling and control scheme. The results indicate that the modeling errors are estimated by the active modeling algorithm with an acceptable estimation error. The tracking error of the AMEC is 52.7% smaller than that of the PID controller. The proposed modeling and control methods for the RPA lay a solid foundation for its biomimetic application.