Real-Time Autonomous Vehicle Navigation via Rule-Based Waypoint Selection and Spline-Guided MPC

This paper presents a robust and efficient Localized Spline-based Path-Planning (LSPP) algorithm aimed at improving autonomous highway navigation. LSPP uniquely combines localized quintic splines with a speed-profile optimizer to generate smooth, dynamically feasible trajectories that prioritize obstacle avoidance, passenger comfort, and strict adherence to road constraints such as lane boundaries. By leveraging real-time data from the vehicle’s sensor fusion module, LSPP accurately interprets the positions of nearby vehicles and obstacles, producing safe paths that are passed to the Model Predictive Control (MPC) module for precise execution. Simulations show LSPP reduces lateral jerk by 30% and computation time by 25% compared to Bézier-based methods, confirming enhanced comfort and efficiency. Extensive testing across diverse highway scenarios further demonstrates LSPP’s superior performance in trajectory smoothness, lane-keeping, and responsiveness over traditional approaches, validating it as a compelling solution for safe, comfortable, and efficient autonomous highway driving.