The Research on PID Control Ways to Improve RBF Neural Network

PID control has been a hot topic and it is a control technology which is widely used in industrial control process. In this paper we researchon PID controller structure and algorithm which based on RBF neural network. It is difficult to determine the clustering radius for the traditional RBF neural network algorithm with the shortcomings of poor stability and low learning efficiency. We improved the nearest clustering algorithm and combine it with the gradient descent to make the mixed learning algorithms and build neural network structure, while adjustthe network parameters and improved PID control method based on RBF neural network. At last we make the verification for its effectiveness and advantages by simulation experiments.