Linearized mathematical model of intelligent pneumatic actuator system: position, force and viscosity controls

Abstract

This paper proposed a linearized model of intelligent pneumatic actuator (IPA) based on the linearization technique known as Taylor series expansion. First, nonlinear mathematical modeling for the IPA system according to fundamental physical derivation is presented. Linearization is then introduced to linearize this nonlinear mathematical model. For the controller design, Neuro-Fuzzy Inference System (ANFIS) is proposed. ANFIS, which combines neural network and fuzzy logic, are adopted and applied to the linear mathematical model to perform position, force and viscosity controls. By training the correct data, membership functions for the fuzzy logic can be obtained through ANFIS toolbox in MATLAB. Closed-loop control for IPA system is conducted and performance the Proportional-Integral (PI) ANFIS controller is analyzed and compared with conventional PI controller. Simulation results show that PI ANFIS controller performed better than conventional PI controller in terms of position, force tracking and viscosity control.