Path Planning for Intelligent Vehicles Based on Improved D* Lite

Abstract

Abstract Typical path planning algorithms are good for static obstacles avoidance, but not for dynamic obstacles, so path planning of intelligent vehicles in uncharted regions is a fundamental and critical problem. This study suggests an improved D* Lite path planning algorithm to address the issues of large corner, node redundancy and close to obstacles in the path planned by D* Lite algorithm. Firstly, in order to increase the safety of the path, the D* Lite algorithm sets the safety distance between the intelligent vehicle and obstacles. Then, the kinematic constraints of intelligent vehicles are introduced to increase the path search direction and avoid path corners exceeding the steering maneuverability of intelligent vehicles. Next, the path is optimized, and the optimization process of removing redundant points is employed to tackle the problem of curved search path and redundant nodes, and the path is smoothed by using third-order Bezier curve to generate a path with continuous curvature. Finally, the enhanced D* Lite algorithm is fused with the improved dynamic window approach (IDWA) to achieve real-time obstacle avoidance based on the global optimal path for moving obstacles. Simulation studies in static and dynamic contexts are used to demonstrate the usefulness of the revised D* Lite algorithm. The results show that compared with other path planning methods, the path generated by the proposed method has more safety and smoothness features, and improves the path quality. Therefore, the proposed algorithm has certain effectiveness and superiority in path planning problems in static and dynamic environments