The research of sampling frequency for A DC servo motor speed control system based on neural networks

This study utilizes the sampling frequency will affect the performance of a direct neural controller (DNC), which is applied to a DC motor speed control system. A direct neural controller of self-tuning strategy is proposed and treated as a speed regulator to keep the motor in constant speed without the specified reference model. A tangent hyperbolic function is used as the activation function, and the back propagation error is approximated by a linear combination of error and error´s differential. The simulation results reveal that the proposed speed regulator keeps motor in constant speed with high convergent speed, but the convergent speed is affected by the sampling frequency. In general, the high sampling frequency will make the speed control system have favor performance, but it will take lots of CPU time. This study applies off-line training to evaluate the appropriate initial values of neural connective weights, then the speed control performance will be improved under low sampling frequency condition.