面向动态避障/路径跟踪任务的大型船舶智能制导与控制策略研究

The collision risk between ships and rocks often exists for marine ship following the waypoints-based planned route. Thus, the developed autopilot system should be with capabilities of the dynamical obstacle avoidance and the path-following functions. For this purpose, a set of new methods are proposed in this project to develop the intelligent autopilot system, where the requirement from the control engineering practice and the uncertain environmental disturbance have been considered. Firstly, the obstacle avoidance maneuvering model is derived by using the mathematics methodology, including three cases: the multi-static obstacles, the multi-moving obstacles and the hybrid multi obstacles. And the on-line path planning is achieved for the obstacle avoidance and path-following tasks, by integrating the security space based optimization algorithm into the DVS (Dynamic Virtual Ship) guidance. Furthermore, the robust neural damping technique and the homogeneous system theory are employed to tackle with the control problem. For constraints from the actuating servo system and the real-time requirement, a novel robust finite-time control algorithm is proposed to obtain the finite-time stability of the corresponding error dynamical system between the ship and DVS. Finally, the effectiveness of the proposal would be verified by the digital simulation and the free-running model based physical experiments. The results of this project can not only enrich the existing marine cybernetics, and provide useful references to facilitate the application of the related theoretical achievements.

船舶自动航行由航路点信息确定计划航线，且不可避免船船、船礁会遇问题，因此要求智能航行系统实现路径跟踪任务的同时，对障碍目标应具有动态避障功能。本项目针对大型船舶动态避障/路径跟踪任务，提出一套考虑控制工程需求和不确定海洋环境条件的船舶智能航行系统设计方法。首先，利用数学演绎方法建立航行水域中多静止/移动目标、混合障碍目标3类情况的避障操纵模型，在动态虚拟船型制导框架下引入安全领域优化算法，实现适用于大型船舶动态避障与路径跟踪任务的实时路径规划；在此基础上，采用鲁棒神经阻尼技术和齐次系统理论相结合，提出一种考虑执行装置约束和计算负载需求的鲁棒有限时间控制算法，保证大型船舶对制导层虚拟船型参考的跟踪达到有限时间稳定；最后通过数字仿真和物理自航模实验验证该技术的可行性和有效性。该项研究不仅可以丰富现有的船舶运动控制理论，而且为智能船舶理论成果在实际工程领域中的应用提供良好的示范和借鉴。

船舶自动航行由航路点信息确定计划航线，且不可避免船船、船礁会遇问题，因此要求智能航行系统实现路径跟踪任务的同时，对障碍目标应具有动态避障功能。本项目针对大型船舶动态避障/路径跟踪任务，深入解析了船舶海上航行面临的大惯性耦合运动、智能化工程任务需求、复杂海洋环境干扰等问题，突破了船舶路径跟踪制导、智能航行避障制导、船舶路径跟踪鲁棒神经网络控制等关键技术创新，构建了系统化的船舶智能航行制导-控制一体化设计方案，取得了一系列围绕智能制导与控制主题的研究成果。本项目共计发表论文40余篇，其中SCI收录32篇，含WOS一区SCI论文27篇；出版专著2部；授权发明专利3项，软著1项；获得海洋工程科学技术二等奖、教育部自然科学二等奖、交通教育科学研究优秀成果奖一等奖、辽宁省研究生教学成果二等奖等重要奖励。研究成果不仅丰富了现有船舶运动控制理论体系，而且为智能船舶理论成果在实际工程领域中的应用提供良好的示范和借鉴。