Two-phrase Optimization Based on Dijkstra and Artificial Potential Field Algorithms for Autonomous Vehicle Routing Problem

Path planning is one of the key technologies in autonomous driving. This paper attempts to combine global and local path planning by proposing the D&APF algorithm, which is based on Dijkstra's algorithm and the Artificial Potential Field method. This algorithm ensures both the safety and comfort of vehicle driving while achieving optimality and smoothness in path selection. Initially, the algorithm extracts nodes and their mutual distances from urban road data, employing Dijkstra's algorithm to select the optimal global path. Then, as the vehicle travels along the path, D&APF constructs a potential field based on real-time road obstacle conditions and employs an improved Artificial Potential Field method for local path planning. Meanwhile, D&APF optimized the attractive and repulsive forces for vehicles, overcoming the failures of traditional Artificial Potential Field methods in specific situations. The paper presents simulation experiments and algorithm comparison on local path planning, with results demonstrating that D&APF can achieve safety and smoothness in path planning along with obstacle avoidance, contributing to the research in the field of autonomous driving.