An Improved Dynamic Window Path Planning Algorithm Using Multi-algorithm Fusion

Aiming at the problems of poor adaptability of traditional dynamic window algorithms and difficult to quickly and effectively plan paths in the face of complex obstacles such as spiral obstacles and narrow obstacles, we propose an improved dynamic windows approach path planning algorithm based on A* algorithm and artificial potential field method fusion. Firstly, we improve the security constraints of the dynamic window algorithm, replace the obstacle distance evaluation function in the original algorithm with the artificial repulsion field function, and add the target endpoint distance sub-evaluation function. Secondly, the improved dynamic window method is integrated with the A* path smoothed by gradient descent method, which solves the problem of poor global planning of the traditional algorithm. And the weight of the evaluation function will adaptively change according to the surrounding environment, which enhanced the adaptability of the algorithm. Finally, through the comparison of simulation results, we verified that the fusion algorithm has a great improvement in planning efficiency, safety, and path smoothness, and is more in line with the motion characteristics of mobile robots.