An Auto-tuning PID Regulator Using Grey Predictor

It is the purpose of this paper to introduce the advantages of grey predictor controllers. We adopt grey prediction to obtain simple and effective estimated values, and, with the aid of first-order low-pass alpha filter, greatly improve the accuracy subsequently used in the prediction for system response. The result will be in turn used to predict error and furthermore automatically adjust the parametrical values of PID controller, and accordingly will be able to deal with the possible variation of system responses at the very first stage. It can not only actively promote the responses efficiency of transient response, but also passively prevent disturbance. As a matter of fact, the highest demand of "plug in and play" can be met without any need to adjust the parameter. This paper will give a detailed specification of the system structure, the design, and the concept, as well as prove the modulation function of grey predictor controllers in unit step response by means of Matlab program simulation and mathematical argumentation. In transient response, it will effectively fasten rising time, shorten settling time, and oppress overshoot; meanwhile, in steady state response, it is able to reduce steady state error to zero and achieve what traditional PID cannot perform