Neural Learning Controller with Tracking Differentiator

It is well known that the closed-loop dynamic quality of the conventional PID correction system is too sensitive to the change of the PID gain value. When the controlled plant is in a changing environment, it is necessary to change the gain according to the plant change in the environment. In addition, when the input signal contains noise, the differentiator in traditional PID control can easily lead to serious noise amplification effect. In view of the above shortcomings, a neuron learning controller with tracking differentiator is proposed, which can reduce the noise of input signal and change the gain of PID according to the change of the environment of the controlled plant. The simulation results show the effectiveness of the algorithm.