Self tuning of PID controller based on simultaneous perturbation stochastic approximation

Abstract

A model need to be identified for the controlled plant when applying advanced strategies to determine the parameters of a PID controller. An online, model-free parameter tuning method is proposed based on simultaneous perturbation stochastic approximation in order to free the control system from plant model. The main idea for this tuning method is to use the highly efficient simultaneous perturbation approximation to the gradient of performance index function for PID controller. All parameters of PID controller can be tuned online by stochastic approximation. Simulation results show that satisfactory performances can be obtained using the proposed tuning method. © 2012 Springer-Verlag.