Key points are not available for this paper at this time.
Aiming to address the defects of the traditional A* algorithm such as mismatch with dynamic environments, insufficient path smoothness and poor real-time obstacle avoidance and the limitations of the traditional DWA (Dynamic Window Approach) algorithm, which easily falls into local optima and relies on parameter tuning, this paper proposes an autonomous driving vehicle navigation path-planning scheme fusing BJA* (Bidirectional Jump point A*) and improved DWA. The fused algorithm enhances A*'s global efficiency via a 24-neighborhood search and a bidirectional jump-point strategy and boosts DWA's local robustness by optimizing the evaluation function and integrating global path information. MATLAB (2022b)-based simulation experiments were conducted. In global path planning, BJA* was compared with improved A* methods, post-processed for path planning, and evaluated through ablation experiments that highlighted the contribution of the improvements; in local obstacle avoidance, vehicle posture, linear and angular velocities under different dynamic scenarios were compared. Experiments show that BJA* exhibits obvious improvements in path length, traversal time and turn number and that integrated local obstacle avoidance makes equipment speed control run more stably.
Xin et al. (Mon,) studied this question.