A hierarchical planning model considering traffic rule compliance and collision avoidance

The existing coupled legality and safety planning problem is complex to deal with, and this paper proposes a hierarchical trajectory planning method with decoupled traffic regulations and motion collision avoidance. The upper layer considers road marking, road orientation, and traffic regulations to establish a feasible candidate set. Then road efficiency, path guidance, and driving burden are considered, and in order to make the above considerations meaningful the recommended speed, desired area, and desired path are determined using collision risk as a rigid requirement. The lower layer treats continuous and discontinuous factors in isolation avoiding the problem of solution discovery. Firstly, the reachable space is determined by making predictions based on the car’s motion state and maneuverability. Then, unsafe trajectories are further eliminated by collision detection algorithms. Then, the optimal solution is obtained by considering path following, speed following, and comfort. Finally, the effectiveness of the methodology of this paper is illustrated by experimental validation in various scenarios using a driving simulator experimental platform.