Performance Improvement of Brushless DC Motor Speed Control Based on FOC Using Reinforcement Learning and PI

Brushless DC (BLDC) motors are widely used in industrial applications, including electric vehicles and robotics, due to their high energy efficiency, reliability, and superior performance. However, controlling BLDC motors under dynamic conditions, where speed setpoints and load torque fluctuate, presents significant challenges. These challenges are particularly evident when using conventional ProportionalIntegral (PI) controllers, which require complex manual tuning. This research proposes a control system for BLDC motors based on Field-Oriented Control (FOC) using a PI controller optimized with Reinforcement Learning (RL). Specifically, the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm enhances the PI controller's performance. The proposed RL-optimized PI controller is compared with a conventional PI controller. Experimental results demonstrate that the proposed method improves speed performance at various load torque by reducing the settling time by 0.32 s, decreasing overshoot by 9.4 %, and reducing steady-state error by 0.66 %, although it slightly increases the rise time by 0.18 s.