This paper describes the principles of BP algorithm and the improved BP neural network be used in the traditional PID control,neural network and PID control law integration,with both self-learning neural networks,adaptive and capacity to approximate any function, Conventional PID control structure also has a simple,high reliability characteristics. Avoid the network into a local minimum;it can speed up the network training speed. So the control can have non-linear,time variability and uncertainty of the complex system of controls.Overcome the PID control parameters of the adjustment process in the system model for over-reliance on the shortcomings.Using MATLAB simulation results show that the BP neural network based self-tuning PID control to control the parameters of the traditional approach to achieve good control effect optimal.
展开▼