A phenomenological model-based controller for position tracking of a pneumatic muscle actuator driven setup

Abstract

Pneumatic muscle actuators (PMAs) have been attracting attention in recent years due to characteristics of natural compliance, high power output, lightweight, and low cost. Various apparatus and robots are developed by adopting PMAs and kinds of advanced controlling strategies are proposed to improve achieving accuracy. However, many of the proposed models and strategies are very complex and not suitable for the actual application. In this paper, we devise a single PMA driven setup and build up empirical quartic polynomial models of the apparatus by fitting sampled pressure-displacement data. The novel models are compact and easily computable. An improved PID algorithm by employing the phenomenological models as feed-forward is devised to improve the dynamical response of the setup. The experiments are performed to assess the proposed controller in improving the position accuracy of the apparatus.