Graph Search Based High-Speed Vehicle Trajectory Generation with Adaptive Node Sampling Based on Patterns for Obstacle Avoidance

Abstract

This paper proposes a high-speed vehicle trajectory generation algorithm based on graph search. The algorithm utilizes driving patterns for obstacle avoidance to limit nodes of the graph without sacrificing obstacle avoidance performance. Experiments were conducted in simulation and the results showed that the proposed algorithm could generate trajectories for driving 15 seconds ahead in average computation time 0.037 second while the rate of avoided obstacles was 100%.