Deep reinforcement learning for permanent magnet synchronous motor speed control systems

Abstract

The permanent magnet synchronous motor (PMSM) servo system is widely applied in many industrial fields due to its unique advantages. In this paper, we study the deep reinforcement learning (DRL) speed control strategy for PMSM servo system, in which exist many disturbances, i.e., load torque and rotational inertia variations. The speed control problem is formulated as a Markov decision process problem, which is computed optimal regulation scheme corresponding to each speed and error state using the deep Q-networks. Simulation results are provided to demonstrate that compared with conventional proportion integral control, the proposed DRL control can improve the robustness against load disturbances and high performance of the PMSM speed control system.