Research on Robot Path Planning Based on Fusion Algorithm of Optimized $\mathrm{A}^{\ast}$ and DWA

Abstract

In the paper, an fusion of optimized A* algorithm and Dynamic Window Approach (DWA) algorithm to pathfinding is presented, combining the strengths of both the A* and DWA algorithms. The innovation of the fusion algorithm lies in two aspects. Firstly, it increases the accuracy of estimating the cost of search nodes by incorporating environment dynamic adjustment parameters <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{e}^{\alpha\cdot \mathrm{k}}$</tex> into the estimation function of A* algorithm. Secondly, it integrates the A* and DWA algorithms, and add a function distB within the evaluation function of DWA algorithm to flexibly avoid unknown obstacles. In the Comparative Experiment, compared to A*, the optimized A* algorithm decreases the path length by 5.54%, the corner count by 58.33%, the sub-node count by 68.18%, and the time by 16.74%. Compared to DWA algorithm, the fusion algorithm reduces the path length by 14.89%, total angel of turns by 53.24%, and the time by 13.06%. The experimental results provided indicate that the fusion algorithm introduced in the paper demonstrates superior pathfinding performance compared to traditional A* and DWA algorithms.