Grey prediction in the application of friction compensation

A method of combining GM (0, N) model with BP neural network is proposed and used in friction compensation. Firstly BP neural networks compensate the friction, then BP network is replaced with GM (0, N) model in a certain moment. With this method the neural network can apperceive the friction torque by online self-learning and compensate. The results of both simulation and experiment show that the method is feasible and effective on the friction compensation, it has well prospect for engineering application.