Automatic Design of Fuzzy Controller for Rotary Inverted Pendulum with Success-History Adaptive Genetic Algorithm

Abstract

In this paper an adaptive genetic algorithm for fuzzy controller design is proposed. The problem of rotary inverted pendulum stabilization and swing-up sequence is considered. The genetic algorithm is used to determine the structure of the fuzzy rule base, which is then used to control the pendulum. The experiments on a model show that the success-history adaptive genetic algorithm is capable of automatic search for an optimal rule base with no prior expert knowledge about the system dynamics.