Hybrid Path Planning Based on Improved A-Star and DWA Algorithms

To solve the traditional A-star algorithm’s problems (redundant node expansion, poor real-time path planning, difficult dynamic obstacle avoidance) in complex dynamic obstacle scenarios, this study fuses DWA with A-star, proposing an A-star-DWA collaborative algorithm integrated with JPS to boost mobile robot path planning efficiency and algorithm convergence. It optimizes path nodes via reconstructing the eight-neighborhood search strategy and embedding JPS’s forced neighbor detection; optimizes A-star’s evaluation function by adding a path-deflection cost function and heuristic weight coefficients; and uses global path critical points as DWA’s dynamic targets, triggering local replanning when obstacle collision distance d ≤ 0.5m. Results show the improved A-star cuts invalid node expansion by 48%, computation time by 63.4%, turns by 57.9%, and path length by 3.2% vs. the traditional one. The fusion algorithm has high dynamic obstacle avoidance success rate and fast post-avoidance path recovery, verifying its engineering value in autonomous driving, industrial logistics and other dynamic environments.