Fuzzy Logic-based Neural Modeling and Robust Control for Robot

Abstract

A fuzzy logic-based neural modeling and robust control method is presented for biped robots. The dynamic inversion of biped robots is used to compensate the complex nonlinearity of the dynamic system and is approximated by the fuzzy neural network. To handle the construction errors of FNN, the H <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</inf> controller is designed to attenuate the effects of the construction errors to a prescribed level. By integrating the H <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</inf> method and the fuzzy logic-based neural modeling technique together, the proposed control method can guarantee the robust performance of the closed loop system. Simulation results show that the proposed method is effective.