面向海域巡逻的多无人船路径规划方法及仿真  

Multiple unmanned surface vehicles-based path planning method for sea patrol

作　　者：王朝飞 王慎执 宋士吉 王凯[3] 武帅 黄高 吴澄[1] Wang Chaofei;Wang Shenzhi;Song Shiji;Wang Kai;Wu Shuai;Huang Gao;Wu Cheng(Department of Automation,Tsinghua University,Beijing 100084,China;Center for Information Research,Academy of Military Sciences,Beijing 100142,China;National Innovation Institute of Defense Technology,Academy of Military Sciences,Beijing 101400,China)

机构地区：[1]清华大学自动化系,北京100084 [2]军事科学院军事科学信息研究中心,北京100142 [3]军事科学院国防科技创新研究院,北京101400

出　　处：《中国图象图形学报》2023年第8期2536-2548,共13页Journal of Image and Graphics

基　　金：广东省科研事业单位重点领域研发计划项目(2020B1111500002);国家自然科学基金项目(62022048);国防科技基础加强计划。

摘　　要：目的随着人工智能技术的大力发展,以无人船为代表的智能无人装备在海洋技术领域已开始实际应用。在大型海域巡逻任务中,相比传统的巡逻装备,多无人船协同执行区域覆盖任务具有运行时间短、效率高以及风险低等优势,但需要解决任务划分、能源补给等难点问题。针对上述问题,提出一套完整的解决大型海域巡逻任务的多无人船协同执行区域覆盖任务的解决方案,并进行面向实际海图的仿真实验。方法首先对多无人船海域巡逻任务进行了问题定义和建模,在构建栅格海域地图的基础上,提出了基于任务均衡的区域划分策略和多种评价指标;在分析比较多种传统遍历算法优劣的基础上,提出了无能源约束和有能源约束条件下的多无人船区域覆盖遍历算法,尤其在有能源约束条件下考虑了多种充电后路径规划策略。结果利用Python搭建仿真应用平台,对三沙市周边15000多平方公里海域进行仿真实验,实验比较了不同遍历算法在无能源约束和有能源约束条件下的优劣,验证了所提出的区域划分策略和遍历算法的有效性和能源约束下无人船在充电后重新规划遍历路径策略的优越性,并对覆盖率和无人船电池里程等参数进行了消融实验。结论提出的方法和仿真分析结果为多无人船协同执行大型海域巡逻任务的实际应用提供了一定的借鉴作用。Objective The artificial intelligence(AI)technology based intelligent unmanned equipments like unmanned surface vehicles have been developing for such domains of marine technology-derived military operation and counterterrorism,resource exploration,seabed topography mapping,water quality testing,and marine search and rescue.In recent years,such issues of unmanned surface vehicles-related path planning have been concerned about.However,existing tasks are still focused on the path planning of single unmanned surface vehicle.The methods and simulations are only based on narrowed water areas like lakes and ports,and the problem of energy constraint is required to be melted into as well.For large-scale sea patrol scenarios,multi-unmanned surface vehicles can be incorporated into coverage path planning tasks in terms of quick response-relevant fast speed,high efficiency and low risk.However,small unmanned surface vessels for sea patrol are mainly driven electrically.Their navigating range and operating time are limited.It is necessary to divide the large-scale sea patrol task into multiple tasks and the problem of energy constraint is required to be resolved for unmanned surface ships.Therefore,we develop a solution of area coverage path planning based on multiple unmanned surface vehicles-integrated large-scale sea patrol tasks.Method First,we define and model the problem of sea patrol of multiunmanned surface vehicles,a task-equal region division strategy and multiple evaluation criteria are proposed and implemented through constructing the sea grid map.The task-equal region division strategy can make all unmanned surface vehicles get similar size of task area and ensure that the full task can be completed to reach the minimum total time.Second,we analyse and compare several traditional traversal algorithms,including round-trip traversal algorithm or spiral traversal algorithm.The breadth-first search algorithm is combined with these traversal algorithms to perform the area coverage path planning task.We demonstrate

关 键 词：海域巡逻 无人船 区域覆盖路径规划 能源约束 仿真 

分 类 号：U675.79[交通运输工程—船舶及航道工程]