Surplus torque suppression research based on the improved wavelet neural network

This paper introduces the working principle and mathematical model of the servo electric load simulator, and puts forward a kind of intelligent controller based on improved wavelet neural network. Bang-Bang controller is used when the error is over the threshold, improved wavelet neural network and fuzzy compensation controller is used when the error is below the threshold; Considering the computational complexity and performance of the control system, the number of hidden neurons is designed. Simulation and experimental results show that the dynamic and static performance meets the target of the phase and the amplitude less than 10°, 10%, respectively. The control strategy significantly improves the system tracking performance and precision of the torque load.