Application of neural network PID controller to elevator control system

The elevator is a kind of complex system with time-varying and strong-coupling characteristics. For elevator systems, with use of traditional PID algorithm, as there are disadvantages of difficult optimal parameters selection, weak steady-state behavior, etc., it is difficult to achieve satisfactory control effect. Therefore, this article discusses the theory of using RBF neural network to identify control object, providing received Jacobian message to BP network, then using arbitrary nonlinear expression ability of BP neural network to achieve the optimum combination of PID control parameters through studying the system, and finally reaching the goal of speedy and stable control. Meanwhile, simulation comparison is made to traditional PID controller on MATLAB and Simulink, and the result shows that the PID controller based on neural networks is faster in response and better in follow nature than the traditional PID controller is.