Identification of Cogging Force in Ironed Linear Motor Based on RBF Neural Networks using Hybrid Self-Adaptive TLBO

The cogging force is an intrinsic characteristic of linear motor, which is caused by the structure of linear motor. As a crucial disturbance, the cogging force seriously impairs the positioning accuracy of linear motor. The existing compensation approaches are hardly implemented or have a low accuracy in practice, due to the inaccurate model of the cogging force. In this paper, the identification problem of the cogging force of the linear motor is investigated. To compensate the unknown nonlinearity, the RBF neural network is utilized for fitting. Moreover, to overcome the local optimal solutions involved with the traditional TLBO method, a hybrid self-adaptive TLBO algorithm is adopted to train the neural network. Finally, experimental results confirm the effectiveness and advantage of the proposed method.