Study on ann-based intelligent self-tuning PID controller

Abstract

PID controllers are widely used in the military field. But the conventional PID controller is difficult to be used for the nonlinear and imprecise mathematical model system, since its parameters can only be obtained from the system with accurate mathematical model. In this paper, the intelligent self-tuning PID controller based on neural network is designed for the system which is nonlinear and does not have mathematical model. The designed neural network of the controller can identify the real-time parameters and the controller can be used in the missile engine hung in the airplane to realize the PID control. The identification is simulated. The experimental results and simulation show that the intelligent self-tuning PID controller based on neural network is feasible. One of the difficulties that design the neural network is to determine the number of the layer, the neurons in the hidden layer and the learning rate. The other is to choose MOBP or SDBP network. To this point, this article has solved the problem for the intelligent self-tuning PID controller using in the missile engine hung in the airplane.