The dynamic fuzzy logic system: Nonlinear system identification and application to robotic manipulators

Abstract

A novel dynamic fuzzy logic system (DFLS) structure is presented, and its properties investigated. Based on the DFLS, an identification scheme is developed using the Lyapunov synthesis approach with a projection algorithm modification, which is shown to be globally stable for a large class of nonlinear dynamic systems in the sense that all system parameters and variables are uniformly bounded. The application of DFLS-based identifier to practical systems is illustrated by identification of a flexible link manipulator with highly nonlinear joint friction. The effectiveness of the identification scheme in dealing with complex systems is demonstrated. The significance of human expertise in improving system performance is illustrated in a simulation experiment. ©1997 by John Wiley & Sons, Inc.