摘要
针对海面上密集多物标会遇场景下无人艇自主避碰问题,综合考虑避碰规则,提出一种基于密度聚类(DBSCAN)与几何避碰算法的物标集群避碰模型(CCAMDG)。在经典的DBSCAN中引入航向差阈值和航速差阈值两个参数,将无人艇周边满足聚类条件的密集动态物标聚类视为需要避让的物标集群,并利用圆形包络算法建立物标集群规避区;在几何避碰算法中引入基于视线的碰撞检测方法,并在《国际海上避碰规则》的约束条件下,实现无人艇对物标集群的安全避碰。仿真结果表明:该方法能够安全地对多动态物标完成避碰行动,同时降低避碰决策复杂度,提高避碰效率,可用于海上避让编队集群、渔船集群等交通流密集情景中的避让决策。
To meet the challenge of autonomous collision avoidance with multiple dynamic targets,Aiming at the problem of autonomous collision avoidance of USV in dense objective encunter scenario on the sea,a novel Cluster Collision Avoidance Model was proposed based on Density-Based Spatial Clustering of Applications with Noise(DBSCAN)and Geometric collision avoidance algorithm(CCAMDG).In this model,two parameters of the heading difference threshold and speed difference threshold were introduced into the classical DBSCAN to deal with the dynamic behavior of targets.The group of multiple dynamic targets clustered by the improved DBSCAN algorithm was regarded as a virtual single dynamic target,which was marked by the circular envelope algorithm for collision avoidance decision.To avoid this cluster of dynamic targets,the geometric collision avoidance algorithm was improved by introducing a new collision detection method to meet the requirements of the Convention on the International Regulations for Preventing Collisions at Sea(COLREGs).The comprehensive simulation results show that the proposed CCAMDG model can safely and effectively complete the collision avoidance actions against multiple dynamic targets,reduce the complexity of collision avoidance decision-making,and improve the efficiency of collision avoidance.
作者
贾世灏
潘家财
魏凯
陆蒙洁
JIA Shi-hao;PAN Jia-cai;WEI Kai;LU Meng-jie(Navigation College,Ji Mei University,Xiamen Fujian 361021,China)
出处
《广州航海学院学报》
2025年第1期33-41,共9页
Journal of Guangzhou Maritime University
基金
福建省自然科学基金项目(2023J01326)
集美大学国家基金培育计划项目(ZP2023004)。
关键词
避碰
密度聚类
几何避碰算法
物标集群
无人艇
collision avoidance
DBSCAN
geometric collision avoidance algorithm
target clusters
unmanned surface vehicle
