A Dynamic Path Planning Algorithm Based on the Improved DWA Algorithm

Because of its superior obstacle avoidance capability, the Dynamic Window Approach (DWA) algorithm has been widely used in local dynamic path planning nowadays. However, in areas with dense obstacles, the DWA algorithm prefers to go around the outside of the dense obstacle area, which increases the total distance. In addition, when encountering a "C" shaped obstacle, the objective cost function will fail and the path will not be found. Therefore, this paper proposes a method to improve the DWA algorithm. Based on the existing constraints, we also propose to score the distance between the current point and the target. In our experiments, we use the traditional DWA algorithm as a reference method and compare the two algorithms in maps with different characteristics. The experimental results demonstrate that the improved DWA algorithm achieves better results in obstacle avoidance.