AIS data-driven MAAC-Stackelberg multi-ship cooperative collision avoidance algorithm

Tie Xu, Tengdong Wang, Jiansen Zhao and Qinyou Hu

PLOS ONE, 2026, vol. 21, issue 6, 1-26

Abstract: Ship collision avoidance has become a focus issue in maritime navigation. Existing methods often struggle to simultaneously meet the hierarchical decision-making requirements of the International Regulations for Preventing Collisions at Sea (COLREGs), address the dynamic uncertainty of ship risk attitudes, and effectively cope with multi-ship coupling risks. To solve the above problems,this paper proposes an algorithm that combines multi-agent systems with game theory, and integrates ship collision avoidance rules into the reward function design. The algorithm constructs a two-stage framework: the risk attitude perception layer uses a Long Short-Term Memory (LSTM) network to predict the short-term motion states of target ships, and dynamically infers the probability distribution of target ships’ risk attitudes through a Bayesian network combined with historical Automatic Identification System (AIS) data and encounter characteristics. The decision-making execution layer integrates Stackelberg game with the Multi-Agent Actor-Critic (MAAC) algorithm, and embeds COLREGs as rigid constraints into the action space to ensure the compliance of the algorithm. Experimental verification is carried out based on historical AIS data and simulation scenarios. The results show that the proposed algorithm has certain advantages in various key indicators,the collision rate, the COLREGs compliance rate, the trajectory smoothness, and the average risk. Statistical significance tests confirm the robustness and superiority of the algorithm. This study provides a reliable technical scheme for ship collision avoidance strategies in multi-ship waters.

Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0345950 (text/html)
https://journals.plos.org/plosone/article/file?id= ... 45950&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pone00:0345950

DOI: 10.1371/journal.pone.0345950

Access Statistics for this article

More articles in PLOS ONE from Public Library of Science
Bibliographic data for series maintained by plosone ().