Research on a local path planning algorithm based on multivehicle collaborative mapping and a potential field method
Chunya Sun,Haixin Jing,Yanqiu Xiao,Guangzhen Cui,Meijie Zhao,Weili Zhang
DOI: https://doi.org/10.1049/itr2.12491
IF: 2.7
2024-03-21
IET Intelligent Transport Systems
Abstract:When automatic‐driving cars perform path planning in a road environment, the sensing range of a single vehicle is limited, resulting in an insufficiently smooth planned path. In response to the aforementioned issues, this paper proposes a path planning method based on multi‐vehicle collaborative mapping in the context of vehicular networks. Through experimental validation, we affirm that our proposed path planning method proactively senses obstacles, effectively models them, and achieves beyond line of sight perception. This method ensures the safe and comfortable guidance of vehicles in completing their path planning. To eliminate blind spots in the field of vision and achieve a safe and collision‐free path, this paper proposes a path planning method based on multivehicle collaborative mapping in the context of vehicle networking. First, a multi vehicle map merging strategy based on the fireworks algorithm is proposed. In this strategy, a dissimilarity objective function based on the concept of grid map similarity is established and an improved fireworks algorithm is used to quickly search for the maximum overlap between local maps, achieving multivehicle collaborative mapping. Second, a real‐time path planning method based on artificial potential field theory is proposed. The information obtained from multivehicle collaborative mapping is first combined with the potential field model to form a multifield coupled road environment model. Then, the obstacle repulsion potential field model is improved to address the issues of traditional artificial potential field methods that target unreachability and poor dynamic response. The feasibility and effectiveness of the collaborative path planning method and single vehicle path planning method are tested through simulation analysis. This paper demonstrates through simulation analysis that the proposed path planning method can effectively achieve beyond line of sight perception and safely and comfortably guide vehicles to complete path planning.
engineering, electrical & electronic,transportation science & technology