A novel adaptive control system of PMSM using RBFNN parallel controller

A novel RBF neural network adaptive parallel controller was proposed for permanent magnet synchronous motor. In this controller, model reference adaptive control (MRAC) and RBF adaptive neural network were combined. MRAC was used to track the reference input, and the effect of the unknown uncertainties and disturbances was overcome by online adaptive tuning network parameters of RBFNN which can implement the mapping between speed error and the control current. It was proved that the presented scheme can guarantee the stability of the designed system and the robustness as well. The simulation results show it works well with high dynamic performance and control precision under the condition of variation of parameters and load torque disturbance.