An Adaptive Trajectory Generation Method for Intersections Without High-Precision Maps Based on Multi-Deciders and Evaluators

This invention proposes a trajectory generation method based on quintic polynomials in the Cartesian coordinate system. The polynomial’s coefficients are determined by boundary conditions, including position, velocity, and acceleration at the start and end points of the curve. The system focuses on generating smooth, continuous trajectories optimized for turning scenarios at intersections, where curvature plays a crucial role. A multi-decision maker and evaluator framework is introduced, which uses virtual speed sampling and curvature evaluations to dynamically generate optimal trajectories. An experience pool stores previous trajectory data to refine decision-making for future route planning.

《一种基于多决策器和评估器的无高精地图十字路口自适应轨迹生成方法》