Incorporating cell map information in fuzzy controller design

Abstract

Cell mapping is a powerful mechanism for evaluating the global behavior of nonlinear dynamical systems. A cell map conveys direct information about system dynamics, including the regions of stability, the set of controllable initial states and system trajectories. This paper shows how cell map information can be incorporated in a genetic algorithm for tuning fuzzy controller parameters. The resulting fuzzy controller exhibits superior performance to controllers designed using conventional techniques. It has near time-optimal characteristics with maximal stability and controllability.<>