Research on Dynamic Collision Avoidance Path Planning Method for Ships

• Published in: International Conference on Automation, Control and Robotics Engineering (Online) pp. 264 - 269
• Main Authors: Zhang, Lei, Wang, Guofeng, Fan, Yunsheng, Chen, Xinyu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 16.07.2025
• Subjects: autonomous collision avoidance; Collision avoidance; COLREGs; dynamic collision avoidance domain; Dynamic Window Approach; Heuristic algorithms; Linear programming; Marine vehicles; Navigation; Path planning; Prediction algorithms; Regulation; Safety
• ISSN: 2997-6278
• DOI: 10.1109/CACRE66141.2025.11119606

In response to the complex challenge of autonomous collision avoidance for intelligent ships operating in dynamic and uncertain maritime environments, this study presents an enhanced local path planning algorithm. The proposed method integrates a dynamic ship collision avoidance domain model, the International Regulations for Preventing Collisions at Sea (COLREGs), and the Dynamic Window Approach (DWA). Specifically, a predictive collision avoidance domain is formulated to anticipate the future positions of surrounding vessels, while a scenario-dependent penalty mechanism-derived from key COLREGs provisions (particularly Articles 14 to 17)-is incorporated to refine the DWA objective function. This integration is designed to improve both the navigational safety and regulatory compliance of autonomous decision-making. Preliminary simulation experiments conducted under representative encounter and crossing scenarios demonstrate the efficacy of the proposed approach, evidencing superior rule adherence and collision avoidance performance compared to the standard DWA algorithm. The findings suggest a viable technical pathway toward achieving safe and regulation-compliant autonomous navigation, though further comprehensive validation across a broader spectrum of operational conditions is warranted.