Conference - Gyeong-Hyeon Kang, Myung-Il Roh, In-Chang Yeo, "A Simulation Method for the Coastal Patrol Mission of Multiple Unmanned Surface Vehicles", Proceedings of ACSMO 2026, Busan, Korea, 2026.05.17-21.

Extra Form
Abstract	With the recent reduction in the available military workforce, the transition toward uncrewed naval ships has emerged as a key priority. Accordingly, technologies enabling USVs (Unmanned Surface Vehicles) to execute missions autonomously have been actively studied, with an increasing demand for multi-USV cooperative operations to overcome the limitations of single-ship deployments. When operating USVs for military purposes, however, it is essential to consider the unique characteristics of naval ships (mission profiles, operational constraints, etc.) that distinguish them from commercial ships, while ensuring reliability and safety through rigorous validation under unpredictable scenarios prior to deployment. In particular, to effectively conduct coastal patrol missions in complex coastal environments, an integrated simulation environment is required to validate the entire operational process, including route planning, collision avoidance, and task allocation. To address this need, we constructed a high-fidelity virtual environment that precisely replicates actual coastal conditions using ENC (Electronic Navigational Chart) data. Based on this environment, we proposed a multi-USV cooperative operation method for coastal patrol missions. First, a formation-constrained route planning method was proposed to ensure the safe navigation of the formation to the target area. Starting from an initial route planned by the A* algorithm, the proposed method iteratively refines the route by shifting waypoints away from land and maritime obstacles, thereby ensuring a safe clearance distance. In segments where the navigable width is insufficient to maintain the formation geometry, the formation is reconfigured by reducing its width, ensuring safe navigation of USVs even in constrained waterways. Next, to achieve explainability and reliability required in military missions, an optimization-based collision avoidance method that integrates VO (Velocity Obstacles) and APF (Artificial Potential Fields) was proposed. By formulating collision risks with static/dynamic obstacles and other ships as penalty terms, each USV optimizes its velocity vector in real time to perform safe and efficient avoidance maneuvers. Finally, upon arrival in the target area, to efficiently search patrol areas comprising multiple islands, an area partitioning and dynamic task allocation algorithm was applied to minimize the overall mission completion time. To validate the effectiveness of the proposed method, various scenario-based simulations were conducted in a virtual environment reflecting the coastal geography of the Republic of Korea. The results show that the proposed cooperative operation method maintains a high level of safety and completes patrol missions even in complex maritime environments.
Publication Date	2026-05-19

Abstract

With the recent reduction in the available military workforce, the transition toward uncrewed naval ships has emerged as a key priority. Accordingly, technologies enabling USVs (Unmanned Surface Vehicles) to execute missions autonomously have been actively studied, with an increasing demand for multi-USV cooperative operations to overcome the limitations of single-ship deployments. When operating USVs for military purposes, however, it is essential to consider the unique characteristics of naval ships (mission profiles, operational constraints, etc.) that distinguish them from commercial ships, while ensuring reliability and safety through rigorous validation under unpredictable scenarios prior to deployment. In particular, to effectively conduct coastal patrol missions in complex coastal environments, an integrated simulation environment is required to validate the entire operational process, including route planning, collision avoidance, and task allocation. To address this need, we constructed a high-fidelity virtual environment that precisely replicates actual coastal conditions using ENC (Electronic Navigational Chart) data. Based on this environment, we proposed a multi-USV cooperative operation method for coastal patrol missions. First, a formation-constrained route planning method was proposed to ensure the safe navigation of the formation to the target area. Starting from an initial route planned by the A* algorithm, the proposed method iteratively refines the route by shifting waypoints away from land and maritime obstacles, thereby ensuring a safe clearance distance. In segments where the navigable width is insufficient to maintain the formation geometry, the formation is reconfigured by reducing its width, ensuring safe navigation of USVs even in constrained waterways. Next, to achieve explainability and reliability required in military missions, an optimization-based collision avoidance method that integrates VO (Velocity Obstacles) and APF (Artificial Potential Fields) was proposed. By formulating collision risks with static/dynamic obstacles and other ships as penalty terms, each USV optimizes its velocity vector in real time to perform safe and efficient avoidance maneuvers. Finally, upon arrival in the target area, to efficiently search patrol areas comprising multiple islands, an area partitioning and dynamic task allocation algorithm was applied to minimize the overall mission completion time. To validate the effectiveness of the proposed method, various scenario-based simulations were conducted in a virtual environment reflecting the coastal geography of the Republic of Korea. The results show that the proposed cooperative operation method maintains a high level of safety and completes patrol missions even in complex maritime environments.

Publication Date

2026-05-19

Gyeong-Hyeon Kang, Myung-Il Roh, In-Chang Yeo, "A Simulation Method for the Coastal Patrol Mission of Multiple Unmanned Surface Vehicles", Proceedings of ACSMO (Asian Congress of Structural and Multidisciplinary Optimization) 2026, Busan, Korea, 2026.05.17-21

Gyeong-Hyeon Kang, Myung-Il Roh, In-Chang Yeo, "A Simulation Method for the Coastal Patrol Mission of Multiple Unmanned Surface Vehicles", Proceedings of ACSMO 2026, Busan, Korea, 2026.05.17-21.

단축키

Articles