面向现代防御系统的多无人机协同优化与决策

Based on game theory and intelligent optimization and decision for multiple unmanned aerial vehicles (UAV) cooperation in complex operational environment and the battlefield situation are studied towards enemy modern defense system, those who have the abilities of concealment, stealth, deception, jamming and the capacity to destroy, etc. Firstly, the issue is analyzed and refined by establishing a coordination control system architecture, so that the practical problems can be abstracted into mathematical models such as threat space models, battlefield situation estimation models, constraint models, game models, task allocation models, route planning models, and so on. Next, the dynamic game of incomplete information structure is used for the confrontational decision problem. Then, intelligent optimization and cooperative game are combined to collaboration task allocation and route planning. Then, receding time horizon optimization strategy is adopted to improve the quality of the optimal solution and thus reduce the time cost in route planning and task allocation. Finally, we will verify these proposed methods through simulation verification mainly and physical verification assisted. This research is more closer to military practice, and thus will provide theoretical guidance for multi-UAVs cooperation in the informational battlefieldis.

本项目采用博弈论和智能优化与决策的方法，针对拥有现代防御系统的对抗敌方，充分考虑敌方的隐蔽、隐身、欺骗、干扰和破坏等能力，在复杂的作战环境和战场态势下，进行多无人机协同的优化与决策研究。首先，针对问题背景，建立协同控制体系结构，将实际问题抽象成相应数学模型，建立诸如威胁空间模型、战场态势估计模型、约束模型、博弈模型、任务分配模型以及航迹规划模型等；其次，采用不完全信息动态博弈方法进行对抗决策；再次，为实现智能决策下的多无人机实时协同，将智能优化与合作博弈相结合，实现任务分配和航迹规划；在此基础上，再运用滚动时域策略来实现任务分配与航迹规划的滚动优化，以提高优化解的质量，减少计算时间；最后，通过仿真验证为主、实物验证为辅，来验证多无人机协同对抗决策与在线任务规划技术。本项目研究更贴近军事实际，将为信息化战场下的多无人机协同提供理论指导。

多无人机协同作战将是未来信息化、网络化战场上的重要模式，为了充分发挥多无人机系统的整体作战效能，项目针对拥有现代防御系统的对抗敌方，考虑其隐蔽、隐身、欺骗、干扰、破坏、追踪、逃逸等自主行为，在复杂的战场环境和不完全信息条件下，从“对抗博弈”的角度研究多无人机系统在协同作战中的优化与决策。项目在多无人机协同对抗的问题建模、博弈决策和在线任务规划方面，提出了不完全信息动态多组对抗博弈模型、捕鱼对策模型及其界栅构造方法、多无人机协同追逃对策的合作机制、风场环境下多无人机航迹规划方法等多项创新性研究成果，在线任务分配与航迹规划的优化时间控制在2秒内，实现了复杂对抗环境下的多无人机协同优化与决策，为对抗性集群博弈决策与任务规划的理论与应用奠定了良好的研究基础。.在国家自然科学基金的资助下，项目组共发表高水平学术论文9篇，其中SCI论文1篇，EI论文8篇，获得授权发明专利1项；另有已录用论文1篇，在审论文2篇（SCI刊源）。