An Adaptive IMC-PID Control Scheme Based on Neural Networks

The conventional internal model control and PID (IMC-PID) provides convenient tuning parameter to adjust the response speed and robustness of the closed-loop system because it has only one tuning parameter. But when the characteristics variation and uncertainty factors are included in the control system, it is difficult to accomplish satisfactory control performance by using conventional IMC-PID controllers. In this paper, we propose an IMC-PID controller combined with a neural network. In this design scheme, NN can get the suitable control parameters of a real-time control system after online learning. The weights of the NN are adjusted by the back propagation method so that the control error can be minimized. Simulation results show the effectiveness of the scheme.