Adaptive Neurofuzzy Sliding Mode Control for Torque Ripple Mitigation in PMSM Drives for Electric Vehicle

Electric vehicles (EVs) require advanced motor control to address challenges like torque ripple, nonlinearities, and variable loads, which can degrade performance and comfort. This paper presents an adaptive neurofuzzy inference system‐based sliding mode current controller (ANFIS‐SMCC) designed for a 2‐kW permanent magnet synchronous motor (PMSM) to improve EV propulsion. By combining ANFIS's adaptive learning, which merges fuzzy logic and neural networks, with the robust sliding mode control, the controller ensures precise d–q‐axis current regulation. It effectively handles PMSM nonlinearities, compensates for 20% stator resistance variations, and maintains stability under dynamic loads across diverse speeds and torques. Built on a detailed mathematical model of EV dynamics and PMSM behavior, the controller was rigorously validated using MATLAB/Simulink simulations and experimental tests on a dSPACE DS1104 platform. It achieves a low torque ripple of 0.5–0.6 Nm (2.0%), a current ripple of 0.35–0.6 A, and a fast 120 microsecond computation time, outperforming traditional ANFIS‐HCC and adaptive neurofuzzy inference system‐model predictive current control (ANFIS‐MPCC) methods with 29%–76% better ripple reduction. Chattering is minimized via saturation functions, enabling reliable real‐time operation within a 5‐μs sampling period. These results position ANFIS‐SMCC as a high‐performance, smooth, efficient, and robust solution for next‐generation EV motor control.