An improved A-star algorithm coupled with graph division and AIS data for ship path planning

• Published in: Ocean engineering Vol. 330; p. 121234
• Main Authors: Yang, Tian-Yu, Zou, Lu, Wu, Zi-Xin, Chen, Wei-Min
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 30.06.2025
• Subjects: AIS data; Generalized Voronoi diagrams; Improved A-star algorithm; Path planning; Traffic feature extraction; AIS data; Generalized Voronoi diagrams; Path planning; Traffic feature extraction; Improved A-star algorithm
• ISSN: 0029-8018
• DOI: 10.1016/j.oceaneng.2025.121234

High computation cost and substantial memory consumption are two significant challenges in the practical applications of the traditional A-star algorithm, particularly for the ship path planning in large-scale navigation areas. To overcome such limitations on the algorithm applicability, a Traffic-Feature Informed A-star (TFIA-star) algorithm is proposed in the study of ship path planning. The regular grid cells for space discretization in the traditional A-star algorithm are replaced with irregular polygons, and the historical ship traffic features in a large area are incorporated. Firstly, the selected Region of Interest (ROI) is initialized by a graph division method based on Generalized Voronoi Diagrams (GVD). Then, the K-means clustering algorithm coupled with the blue noise sampling is adopted as the traffic feature extraction method to incorporate ship traffic patterns from historical Automatic Identification System (AIS) data into path planning. To evaluate the performance of the proposed TFIA-star algorithm, simulations of two representative cases with both simple and complex navigation scenarios are conducted. The path planning results and the analyses in terms of five evaluation criteria indicate that, compared with the other path planning algorithms, the proposed TFIA-star algorithm achieves superior path planning performances with smaller computation costs and less memory usages. •A-star Algorithm is improved to reduce computation time and memory usage for large-scale problems.•Regular grid cells are replaced with irregular polygons by Generalized Voronoi Diagrams (GVD).•GVD coupled with blue noise sampling is applied to initialize the navigation map of path planning.•Ship traffic features extracted from AIS data are integrated into TFIA-star algorithm.•TFIA-star is superior to the other algorithms with better path-planning effectiveness and robustness.