Comparison of two neural network methods applied to a tracking control system

This paper presents the comparison of the backpropagation neural network and the random research method applied to the tracking control of a direct current motor. The neural network control system uses a very simple scheme and it can use in real-time. This technique is based on an adaptive distributed architecture. The comparison of the two alternatives is done in the learning process and the performance of the control system. The neural network is applied as a tracking controller, with parameter error trajectories inputs. The controller classifies the feedback error signals and generates the appropriate control action to the motor. The control system can follow any arbitrary trajectory, even when the trajectory is changed to that not used in training process. The neural network is programmed in a personal computer. The motor speed is sensed with a tachometer, and the output of the neural network actuates on the field voltage of the motor