面向巡航及搜救的群海事无人艇路径规划研究

Maritime unmanned surface vehicles excel in the flexible control ability, excellent autonomy, by using which they can fulfill the marine management and rescue assignments under all-weather condition with ease. To respond to the strategy of becoming a maritime power and the safety of maritime security, the concepts and facilities of the swarm of maritime unmanned surface vehicles (SoM) are introduced into the field of maritime patrol, search and rescue. To assure the successful compliments of the maritime emergency tasks, the high-quality paths should be achieved for the swarm of maritime unmanned surface vehicles. Thus, we focus on the path planning problems for the swarm of maritime unmanned surface vehicles which are oriented to the maritime patrol, search and rescue. Firstly, when comes to the formulation process, the complex environment is merged from the perspective of path planning, including the effects brought by the moving and static obstacles, and the waterborne navigational regulations are performed in the constraint optimization method, the probability velocity obstacle is used to formulate the disturbance of the wind and wave, and the uniform current is also taken into account. Following that, the backgrounds of the maritime patrol and rescue are considered in the modeling process of our path planning problem, the dynamic searched targets are predicted by way of the multi-step combined Runge-Kutta method. Thirdly, several optimal objectives, including the minimized the summation of the patrol or search and rescue path lengths and consumed time, and the maximum benefits of the patrol or the search and rescue are proposed in the formulated multi-constraint time-varying nonlinear multi-objective path planning model. After probing into the above model, the multi-objective gravity search algorithm with the strategy of non-dominated sorting is put forward. Within the proposed algorithm, the Pareto frontier approximation can be enhanced and the ability of searching the diversity resolutions can be improved generously. Finally, to verify the effectiveness of our formulated path planning model and the novel approach, we tend to construct the swarm of maritime unmanned surface vehicle platforms in the form of the simulation and the reduced scale forms. A wide range of simulation experiments under quantities of patrol, search and rescue scenarios is presented. Meanwhile, generous waters experiments would be carried out by using the reduced scale platform. The concerned studies would enrich the content of the maritime security. Furthermore, the formulated path planning model and algorithm would do great deeds for the swarm of maritime unmanned surface vehicles path planning not only with the guidance of theory but also the technical support.

海事水面无人艇操纵灵活、自主性强、能实现全天候的海事管理及搜救。本项目立足国家海洋强国战略及海事安全保障重大需求，在海事保障领域中引入群海事无人艇(SoM)，研究面向巡航及搜救的SoM路径规划问题，为其快速响应监管及营救任务提供高质量的路径。本项研究从路径规划层面融合复杂水域环境，基于约束关系式表征障碍物及航行规章约束、概率速度障碍物表征风浪，考虑均匀流影响；融合巡航及搜救背景，基于多步复合龙格库塔法预测动态搜救目标；以最小化总巡航/搜救航程、最小化总巡航/搜救时间、最大化巡航/搜救效益等为优化目标，构建“多约束时变非线性多目标优化”路径规划模型；探析模型特质，提出非支配排序的多目标引力搜索算法，以增进帕累托前沿逼近程度和搜寻良好多样性解的能力；搭建SoM平台，开展仿真实验及水域试验，验证模型及算法的有效性。本项研究丰富海事保障内涵，凝练的模型及算法为SoM路径规划提供理论指导和技术支撑。

随着人工智能技术的全面兴起和应用，船舶工业迎来了新一轮的发展契机，以无人船艇为代表的自主航行船艇在海事管理、水上科考、渔业监管等方面发挥着越来越重要的作用。无人船艇操纵灵活、自主性强，能实现全天候的海事管理及搜救。本项目立足国家海洋强国战略及海事安全保障重大需求，率先在海事保障领域中引入群海事无人艇(SoM)，研究面向巡航及搜救的SoM路径规划问题，为其快速响应监管及营救任务提供高质量的路径。项目研究成果丰硕，圆满完成既定任务目标。.项目主要创新为：(1)针对海事安全保障任务，开展洋流扰动下海事无人船艇多目标路径规划研究，为突破计算维度的限制、鲜有考虑3个以上时变非线性优化目标路径规划的难题，提出一种动态增强多目标粒子群优化算法，结果证实了所设计算法的优越性能；(2)针对风浪作用下的海事无人船艇巡航路径规划问题，考虑多个巡航点和障碍物下的约束，将其表征成带有非欧式距离的旅行商问题，建立海事无人船艇巡航路径规划模型，提出一种两阶段法，案例证实了上述模型及算法的有效性；(3)针对群海事无人船艇SoM避碰问题，鉴于深度强化学习算法在动态环境下可以表征和控制极为复杂系统的能力，建立基于深度强化学习的智能避碰模型，由决策神经网络给出一系列避碰决策，实现多艘船艇自主安全避碰；(4)自主研发系列面向巡航及搜救的海事无人船艇智能作业平台、风帆辅助驱动无人船艇智能控制系统。.项目研究成果发表在国内外权威期刊上，包括IEEE TII, OE, ISA T, IEEE TNNLS等国际刊物上9篇SCI论文、中国航海等国内知名期刊上多篇论文；出版专著1部；授权专利9项；主持获得省部级科技成果奖励二等奖1项；荣获省部级荣誉称号1项；开展1年及以上国际交流合作3项，开展国内外项目交流报告8次；培养海事保障方向博士生1名、硕士生4名；指导研究生获得国家奖学金2人次、湖北省优秀学士论文1篇、两岸三地规模最大的全国海洋航行器设计与制作大赛5项。.本项研究为推动成果落地搭建了SoM平台，为群无人船艇应用于海事保障领域提供了强有力的理论指导和技术支撑，有效地丰富了群海事无人船艇路径规划理论技术及其应用，实质性地拓展了智能海事保障领域内涵。