Research on path planning of self-driving vehicles based on improved DWA algorithm

Abstract

Abstract With the increasing degree of vehicle intelligence, unmanned vehicles have been widely used in civilian and military fields, which is of research value. In this paper, considering the vehicle dynamics constraints and the efficiency of the algorithm calculation to improve the DWA algorithm for the path of the self-driving vehicle, taking the intelligent vehicle node as the center of the circle and the sensor detection distance as the radius to extract the local map, combining the improved A* algorithm to fuse the DWA algorithm to carry out the local path planning at the key points, and through the design of the simulation experiments in the static and dynamic environments, the results show that, in the static environment of the simulation experiments The results show that in the static environment, the changes of the swing angle of the three vehicles fluctuate between -3 and 4, and the changes of the lateral speed of the vehicles fluctuate between -0.06 and 0.08, which are within a reasonable range of changes and satisfy the safety requirements. In the dynamic environment simulation experiments, the curve amplitude of the curvature similarity evaluation function in the improved algorithm is 40% less than that in the traditional algorithm, the number of iterations is 240 times less, and the car can reach the end point faster. This research can improve path planning accuracy, dynamic obstacle avoidance ability, and better path planning effect, which can be applied in the field of intelligent vehicles.