Chaotic Self-Tuning PID Controller Based on Fuzzy Wavelet Neural Network Model

Abstract

PID controller has been quite successful when its parameters are tuned properly. However, it fails in varying situations. This is even more critical where the system is not known. In this paper, after illustrating the capability of fuzzy wavelet neural network (FWNN) in modeling of nonlinear systems, a self-tuning PID controller based on this model has been designed. The auto-tuner effectively handles the limitation of PID controller in unpredictable conditions such as environmental changes. Chaotic optimization method which is a robust algorithm of escaping from local minimum is applied for tuning of the controller parameters. This supports finding optimal values of controller parameters in a short time which enables online implementation. Unlike most of the researches in this area, the tuning rules could be begun without any trial and error. Since it contains no extra parameters, it has the feature of simplicity in the tuning. The proposed method with few parameters has the ability to increase the speed of tracking with very little steady-state error. The capability of FWNN in the modeling of nonlinear systems with a few rules and susceptibility of the proposed controller will be shown by simulation.