IAPF-RRT*: An Efficient Iterative Optimization Path Planning Algorithm for Autonomous Vehicles
Menglun Su, Chao Yi Wei, Jibin Hu, Qing Huang, Mengjie Zhang, Zhong Kang · 2023
Path planning module is an important part of autonomous vehicles, and its efficiency directly affects the steering safety. Sampling based algorithm has been widely used in autonomous vehicles, and the combination of RRT algorithm and artificial potential field method (APF -RRT*) has been proved to greatly improve the efficiency of planning. However, when driving in long distance and complex scenarios, there are still a lot of repetitive work during the cycles, so this paper proposes an iterative APF-RRT* (IAPF-RRT*) algorithm, which reconstructs the historical path after pruning as a prior search tree of the current frame and uses the nearest neighbor method to retrieve nodes. Considering that the results of the RRT* algorithm cannot be directly applied to autonomous vehicles, this paper takes the results of the IAPF-RRT* algorithm as the initial solution for quadratic programming. The experimental results show that the algorithm has superior real-time performance and the trajectory quality is sufficient to be directly applied to automatic driving.