Neural Sliding Mode Control for Brushless DC Motors

The approach which combined sliding mode control and neural networks is researched for the position controller of brushless DC motors in industry. A new neural sliding mode control scheme was proposed for reducing chattering of sliding mode control in the paper. A global sliding mode manifold was designed in this approach,which guarantees that the system states can be on the sliding mode manifold at initial time and the system robustness can be increased. A radial basis function neural network( RBFNN) was applied to learn the maximum of unknown loads and external disturbances. Based on the neural networks,the switching control parameters of sliding mode control can be adaptively adjusted with uncertain external disturbances and unknown loads. Therefore,the chattering of the sliding mode controller was reduced. The simulation results prove that this control scheme is valid by simulation experiments.