Intelligent Longitudinal Control for PMSM-Based Autonomous Vehicles Using YOLOv10-S and a Fuzzy-PI Controller

This study presents an intelligent longitudinal control system for autonomous electric vehicles using permanent magnet synchronous motors (PMSMs). The system integrates real-time traffic sign recognition with advanced control strategies to improve speed regulation, energy efficiency, and ride comfort. A lightweight YOLOv10-S model is trained to accurately and quickly detect speed limit signs under diverse conditions. The detected speed is used as a reference input for a fuzzy PI controller designed using two approaches: a Linear Quadratic Regulator (LQR) and a robust H∞ strategy based on a Linear Integral Lyapunov Function (LILF) via Linear Matrix Inequalities (LMI). Simulation results highlight that the YOLOv10-S model achieves high accuracy with low detection losses and strong mAP scores across IoU thresholds (0.5–0.95). The proposed control system demonstrates clear advantages over traditional PID and LQR controllers. The LILF-based controller achieves a rise time of 5 ms, zero overshoot, zero steady-state error, and complete rejection of load-induced vibrations. Meanwhile, the LQR controller ensures energy-efficient operation with a 9 ms rise time and reduced current consumption. By integrating deep learning perception with robust control techniques, this system enables autonomous vehicles to dynamically adapt to traffic regulations, resulting in more accurate speed tracking, smoother driving, and lower energy consumption, key benefits for real-world electric mobility. © 2025 The Authors.