针对目前尚未形成一套严谨的船舶智能避碰决策算法完备性验证方法，基于船舶碰撞危险度动态评估体系，提出船舶拟人智能避碰决策（PIDVCA）算法的仿真测试目标，确保在追越局面下智能避碰的安全性和可靠性。通过船舶智能操控仿真测试平台，对船舶拟人智能避碰决策算法进行仿真测试，采用双智能船进行不同交汇特征、不同危险等级以及不同船型的全面深度测试。重点分析双智能船的典型场景及特殊场景测试结果，并对算法中的问题进行优化。优化后的算法在追越局面中能够提供符合《国际海上避碰规则》要求的避碰决策，同时体现海员的通常做法及良好船艺。研究成果为算法的实用化测试认证提供了新模式，并为算法完备性测试和场景库设计提供了新思路。
In view of the lack of a rigorous verification method for the completeness of vessel intelligent collision avoidance decision algorithm, based on the dynamic assessment system of ship collision risk, a simulation test objective of personifying intelligent decision-making for vessel collision avoidance（PIDVCA）algorithm is proposed to ensure the safety and reliability of intelligent collision avoidance in overtaking situation. Through the Ship Intelligent Handle and Control simulation test platform, the simulation test of PIDVCA algorithm is carried out, and the comprehensive depth test of different intersection characteristics, different danger levels and different ship types is carried out by using two intelligent ships. The test results of typical scenarios and special scenarios of the dual intelligent ship are analyzed, and the problems in the algorithm are optimized. The optimized algorithm can provide collision avoidance decisions in line with the requirements of the International Regulations for Preventing Collisions at Sea (COLREGs) in overtaking situation, while reflecting the ordinary practice of seamen and good seamanship. The research results provide a new model for the practical testing and authentication of the algorithm, and a new idea for the algorithm completeness testing and scenario library design.
2025,47(19): 168-175 收稿日期：2024-10-27
DOI：10.3404/j.issn.1672-7649.2025.19.027
分类号：U675.96
基金项目：国家自然科学基金资助项目（51879119）；工信部高新技术项目（MC-201920-X01）
作者简介：艾小波(1982-),男,大副/讲师,研究方向为交通信息工程及控制
参考文献：
[1] 孙峰, 蔡玉良, 马吉林. 船舶智能避碰策略测试方法与指标研究[J]. 交通信息与安全, 2019, 37(5): 84-93.
SUN F, CAI Y L, MA J L. A study of test methods and indicators for marine intelligent anti-collision strategy[J]. Journal of Transport Information and Safety, 2019, 37(5): 84-93.
[2] WOERNER K, BENJAMIN R M, NOVITZKY M, et al. Quantifying protocol evaluation for autonomous collision avoidance[J]. Autonomous Robots, 2019, 43(4): 967-991.
[3] LØVOLL E. Evaluating collision avoidance algorithms in urban and semi-restricted waters using fuzzy logic[D]. NTNU, 2022.
[4] 李丽娜, 杨神化, 尹勇. 船舶自动避碰仿真平台的构建与测试方法研究[J]. 中国航海, 2006(3): 47-50.
LI L N, YANG S H, YIN Y. Study of simulation platform construction for automatic ship collision avoidance and its test method[J]. Navigation of China, 2006(3): 47-50.
[5] 尹勇, 张秀凤, 张显库. 船舶自动舵和自动避碰算法仿真测试平台的研究[J]. 系统仿真学报, 2007, 19(1): 56-58.
YIN Y, ZHANG X F, ZHANG X K. Study on simulation platform for testing algorithms of autopilot and automatic collision avoidance[J]. Journal of System Simulation, 2007, 19(1): 56-58.
[6] YANG S , LI L , SUO Y, et al. Study on construction of simulation platform for vessel automatic anti-collision and its test method[C]//Automation and Logistics, 2007 IEEE International Conference on. IEEE, 2007.
[7] CHEN G Q , YIN Y , LI L N, et al. New design of intelligent navigational simulator[J]. Applied Mechanics & Materials, 2011, 97-98: 896-902.
[8] LI L , YANG S , SUO Y, et al. Automation method for personifying intelligent decision-making for vessel collision avoidance[C]//IEEE International Conference on Automation & Logistics. IEEE, 2008.
[9] CHEN G . The PIDVCA method in the application of intelligent navigation simulator[J]. Navigation, 2016: 1-13.
[10] 李丽娜, 陈国权, 杨凌波, 等. 船舶拟人智能避碰决策算法测试及应用[J]. 中国航海, 2022, 45(1): 1-7.
LI L N, CHEN G Q, YANG L B, et al. Test and application of personifying intelligent decision-making algorithm for vessel collision avoidance[J]. Navigation of China, 2022, 45(1): 1-7.
[11] 李丽娜, 陈国权, 马吉林, 等. 船舶拟人智能避碰决策[M]. 北京: 海洋出版社, 2023.
[12] 李丽娜, 陈国权, 李国定, 等. 一种关联避让行为的船舶碰撞危险及危险度评价方法: CN109887339A[P]. 2020-9-15.
[13] 翁建军, 刘明俊. 船舶值班与避碰[M]. 武汉: 武汉理工大学出版社, 2016.
