A performance analysis of various neuro-fuzzy approaches for controlling nonlinear systems

Nonlinear systems are becoming an area of great interest in the control engineering community. Many interesting problems such as controllability, input-output decoupling, feedback linearization, have been approached with success. On the other hand, not so many results have been achieved in the solution of the problem of identification and control of unknown nonlinear systems. The application of linear methods to nonlinear systems is not very successful when wide control ranges are necessary. For those cases non-conventional methods, such as the use of neural networks, have been investigated. Another promising approach is the application of fuzzy logic to the control of some classes of nonlinear systems. The method is simple, not time consuming, and requires little knowledge of the system equations. The combination of these two methods have been tried with success and is known as a neuro-fuzzy system. This paper presents an overview of the various neuro-fuzzy approaches, and their application to the control of nonlinear systems