复杂海况下的海事无人艇路径规划研究

Maritime unmanned surface vehicle (MUSV) is characterized by agilely control ability, strong autonomy, and powerful maritime management and rescue function under all-weather condition. It demonstrates that MUSV can do great deeds in the field of maritime security. To deal with the unknowns and fulfill the task of maritime supervision and rescue in time, the high quality planned paths should be planned for MUSV. We aim at planning scientific, effective and safe paths for MUSV which would be in accordance with COLREGS and under complex sea condition, such as experienced the gale wind, abrupt wave and current. Within our study, COLREGS is fused based on the method of constraint optimization, the disturbance brought by the complex sea condition is taken into account by the approach of probability obstacles, and the multi-objective including minimized the summation of path length and travel time are proposed in our multi-constraint time-varying multi-objective model. When the path planning model for MUSV has been put forward, the improved multi-objective particle swarm optimization algorithm is devised to solve the path planning model. The newly algorithm adopts the relation of Pareto dominations to update particle extreme, and provides all of the extremum value from non-inferior solution of the archival. Then, the effectiveness of our improved approach is verified by the simulation experiments under generous complex sea condition. We would propose one novel path planning mode and the exquisite algorithm for MUSV path planning. The research targeted will resolve path planning problems for MUSV under complex sea condition evidently.

海事无人艇具有操纵灵活、自主性强、能实现全天候的海事管理和海上搜救等优点，在海事保障领域应用前景广阔。为应对各种未知事件，实现快速响应海事监管、营救等功能，海事无人艇就必须依靠高质量的规划路径作为保障。本项目研究复杂海况下的海事无人艇路径规划问题，在遵循国际海上避碰规则前提下，考虑风、流、浪等复杂海况扰动，规划出科学合理的路径。本项研究基于约束最优化实现避碰规则融合，采用概率障碍物等方法整合复杂海况的扰动，以最小化总航程、最小化总时间等为优化目标，建立"多约束的时变多目标非线性最优化"路径规划模型；分析模型特征，提出改进的多目标粒子群优化算法，该算法采用Pareto支配关系更新粒子极值，应用精英归档及技术，由档案库的非劣解提供全部极值；多海况下开展仿真实验，验证算法的有效性。通过研究，凝炼出一种具有复杂约束的时变多目标路径规划模型及智能求解算法，为海事无人艇路径规划提供理论指导和技术支撑。

海事无人艇在海事监管、应急搜救中发挥着重要作用，高质量的规划路径是其实施各项活动的关键要素。本项目关注复杂海况下的海事无人艇路径规划问题；考虑风浪流等环境扰动，以最小化总航程、总航时等为优化目标，构建多约束时变非线性多目标优化的路径规划模型；设计并实现一种行之有效的增强多目标粒子群优化算法；多场景仿真案例证实了路径规划模型及智能求解算法的有效性。搭建2艘缩尺度无人艇，在长江荆州段、湖泊等水域开展试验工作，验证了本项目模型及增强多目标粒子群优化算法的实用性及有效性。本项研究凝炼出一种具有复杂约束的时变多目标路径规划模型及智能求解算法，开展水域路径规划试验，为海事无人艇路径规划提供理论指导和技术支撑。