Artificial Intelligence DTC MTPA Strategy Based on Speed MRAS Observer for Electric Vehicle Traction Applications

Abstract

This paper presents an advanced direct torque control (DTC) strategy incorporating artificial intelligence and a speed Model Reference Adaptive System (MRAS) observer for permanent magnet synchronous motors (PMSMs) used in electric vehicle traction. The studied electric vehicle is equipped with four in-wheel PMSMs, requiring an electric differential to ensure balanced torque distribution and vehicle stability, especially during cornering maneuvers. To reduce system weight and enhance efficiency, two machines on the same vehicle side are powered by a single three-leg inverter, forming a multi-machine single-inverter configuration. A master–slave control structure is adopted to manage this architecture and ensure synchronized operation of all motors. The proposed AI-based DTC combined with the MRAS speed observer significantly improves torque accuracy, dynamic response, and robustness against disturbances. Simulation results obtained using MATLAB/Simulink confirm that the proposed strategy achieves high performance in both transient and steady-state conditions, ensuring reliable traction, enhanced vehicle stability, and improved overall dynamic behavior.