Abstract—This paper presents a hybrid control framework for autonomous vehicles, combining semantic lane detection with a two-tier control approach: Nonlinear Model Predictive Control (NMPC) for lateral trajectory tracking and Fuzzy Proportional–Integral–Derivative (Fuzzy PID) control for longitudinal velocity management. Real-time visual data from the UltraFast segmentation network is integrated into the NMPC optimization, improving road boundary tracking in dynamic conditions. The fuzzy PID controller is optimally tuned and enhanced with a feedforward compensation branch to anticipate velocity changes, speeding up convergence while ensuring stability. Simulations across various velocity targets demonstrate rapid convergence, lateral stability, and reduced control effort. Compared to classical methods like Linear Quadratic Regulator (LQR) and Pure Pursuit (PP), the proposed system achieves superior tracking accuracy, robustness, and smoother control. Key contributions include incorporating UltraFast-based Lane segmentation into NMPC and using feedforward-enhanced Fuzzy PID for better speed regulation, offering a scalable and adaptive solution for intelligent vehicle control in structured settings.

Keywords—vehicle control, Nonlinear Model Predictive Control (NMPC), Fuzzy Proportional Integral Derivative (Fuzzy PID), vision-based navigate

Cite: Thanh-Ha Vo, Huu-Giang Nguyen, and Hong-Quang Nguyen, "Trajectory Tracking for Autonomous Vehicles Using NMPC Method and Semantic Lane Inference," International Journal of Mechanical Engineering and Robotics Research, Vol. 14, No. 6, pp. 617-629, 2025. doi: 10.18178/ijmerr.14.6.617-629

Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).