DESIGN OF OBSERVER BASED ADAPTIVE PID CONTROLLER FOR NONLINEAR SYSTEMS

摘要

An observer based adaptive PID controller with micro genetic algorithm (MGA) tuning its parameters is proposed in this paper. The task of the controller is to track the desired trajectory of a nonlinear system as best as it could. The proposed adaptive PID controller consists of 3 components including the PID controller, the control compensating for the external disturbance and the modeling error, and the supervisory control. The parameters of PID controller are learned using MGA. The Lyapunov stability theorem is utilized to design the fitness function for MGA so that the parameters under learning by MGA tend to stabilize the nonlinear system. To cope with the situation that the states are not always observable, an observer is designed and integrated with the adaptive PID controller.