Identification of 2-DOF pneumatic artificial muscle system with multilayer fuzzy logic and differential evolution algorithm

Abstract

This paper proposes a new method for identifying a nonlinear pneumatic artificial muscle (PAM) 2-dof parallel system based on the novel NARX multilayer fuzzy model optimized by differential evolution (DE) algorithm. A multilayer fuzzy system is created by combining several MISO multilayer fuzzy models. Each MISO multilayer Fuzzy model is implemented through several Fuzzy Takagi-Sugeno sets. Then fuzzy structures and fuzzy rules of proposed multilayer fuzzy model were optimally trained by DE algorithm. The experiment results are presented. It proves a promisingly scalable and simple method to successfully identify nonlinear MIMO system.