Actor-Critic Algorithm Based R-L Controller for Performance Improvement of BLDC Motor Drive Train

Now a days BLDC motor is used for electric vehicle drive train applications. A reinforcement learning based Actor-Critic model algorithm with Twin-Delayed-Deep-Deterministic (ACR-TD3) agent approach proposed for dynamic performance improvement of BLDC motor drive train under various operating conditions. The novelty of the proposed Actor-Cretic R-L controller is implementation of machine learning algorthem in the control of BLDC motor drive train. This proposed controller observes the error states rigorously and maps to target states to obtain reference speed of BLDC drive train under different operating conditions. The proposed Actor-Critic model based algorithm withTD3 tested on the BLDC drive train and the performance of Actor-Critic reinforcement model with TD3 agent compared to FOPI and PI controllers under various modes of operation using MATLAB-Simulink.