Identification and Control of a DC Motor System based on Neural Networks

Jinzhu Peng and Rickey Dubay


Neural network, System identification, PID control, DC motor


An adaptive control approach based on neural networks is presented to control a DC motor system with friction. Two types of neural network are proposed to facilitate the adaptive control approach. A dynamic neural network (DNN) is proposed to formulate the Hammerstein model, which comprises of a nonlinear static block and a linear dynamic block. And a feedforward PID-type neural network (PIDNN) is presented to formulate the traditional PID controller. The DC motor system is identified by the DNN identifier, which provides system information to facilitate adaptive PIDNN control. Simulation on a DC motor system demonstrated the proposed identification and control performances.

Important Links:

Go Back