Indirect adaptive neural controller of nonlinear systems using auto-tuning neuron

In this paper, a novel indirect adaptive neural controller using only two auto-tuning neurons is developed for a class of nonlinear systems. Unlike traditional multi-layered neural controllers, the structure of the proposed controller is very simple and practicable. There are three adjustable parameters in each auto-tuning neuron. Two such auto-tuning neurons used in our proposed indirect adaptive controller are used to track on-line the desired signal. The adaptation law for adjusting these parameters is developed based on the Lyapunov approach. Moreover, the stability of the overall closed-loop system can be analyzed and guaranteed by introducing the additional supervisory controller and the technique of modified adaptation law with projection. Finally, the tracking control of the inverted pendulum system is presented to illustrate the proposed method.