Empirical Modeling and Position Control of Single Pneumatic Artificial Muscle

Abstract

The model of pneumatic artificial muscles (PAMs) is the foundation of their control. Compared to theoretical model, empirical model is based on experiments and doesn’t need accurate understanding of PAMs. Existing ones have not considered the history-dependence of hysteresis within PAMs which causes creep and deteriorates the controller behavior. In this paper, a novel synthetically empirical model of PAMs which consists of hysteresis element, viscous damping element, rubber elasticity element and contractile element in parallel is proposed. The accuracy of the empirical model is verified by experiments. This model is then applied into the position control of single PAM using fuzzy logic. To overcome the negative impact of hysteresis and to improve the response speed of the controller, the proposed model is then used to calculate an on-line feedforward voltage to the solenoid valve for compensation. In order to improve the robust of the controller, a self-organizing mechanism is introduced. Experimental results indicate that the proposed controller effectively alleviates the creep of PAM and improves in response speed and robustness while maintaining stability.