毕业生知识库

大规模无人机协同目标覆盖是无人机协同决策领域的一个重要研究方向，在军民领域均有广泛的应用需求和重要的理论研究意义。本文以大规模无人机在动态环境中协同对地目标探测并覆盖为背景，针对仅有局部信息的条件下，无人机群保持通信连通并优化覆盖动态目标的协同决策问题。以图论和博弈理论为基础，开展系统性的问题建模、控制架构研究、优化算法设计、以及实物验证平台搭建并验证工作，主要工作及贡献包括:

（1）针对大规模无人机协同目标覆盖问题，深入分析并建立了数学优化模型。同时，结合相关博弈理论知识，证明了大规模无人机协同目标覆盖是一个精确势博弈过程，从而存在纳什均衡解，为后续方案设计提供了理论指导。

（2）提出了扁平式控制架构下的大规模无人机协同目标覆盖方法，该控制架构主要面向任务区域较小的任务场景中。首先，为了降低算法计算复杂度，将地图进行蜂窝栅格化，同时设计目标分组机制以简化目标分配过程。其次，基于上述势博弈的理论指导，结合基于共识的拍卖算法设计了分布式局部协商算法，以实现分布式控制下保持通信连通的同时优化覆盖地面目标。最后，通过仿真测试对该方法的有效性进行测试。

（3）提出了分层控制架构下的大规模无人机协同目标覆盖方法，该控制架构主要面向任务区域跨度大，目标数量多的任务场景中。首先，针对扁平式控制中资源调度效率低下问题，设计顶层中央处理器，负责全局资源管理与调度工作。其次，为提高无人机协同决策效率，对目标进行聚类分组。针对精确势博弈的大规模无人机协同目标覆盖问题，设计了推荐与选择分布式算法。该算法不需要进行多轮协商迭代就可以做出决策，各无人机以最大化自身效用函数来求取全局次优解。再次，针对大规模无人机执行任务过程中通信量巨大，导致通信爆炸问题，设计了态势信息理解与萃取机制，使得大规模无人机系统有更好的可扩展性与更高的任务执行效率。最后，通过仿真测试对该方法的有效性进行测试。

（4）设计并搭建了实物验证平台，并针对以上两种目标覆盖方法进行有效性验证。首先，完成实验平台的组网通信架构及硬件系统架构设计。其次，搭建了半物理算法验证平台，对扁平式控制架构进行半物理实验验证，通过多组实验测试验证了算法的有效性。最后，基于半物理实验的成功测试，将板载计算机搭载至无人机/车上完成实物测试，表明本文提出的分层控制方法在大规模无人机协同目标覆盖中具有很好的目标覆盖性能与可扩展性。

综上所述，本文分别从理论分析、系统控制架构、目标优化分配及通信连同保持方面系统研究了大规模无人机协同目标覆盖问题，并开展了数字仿真、半实物及实物试验验证，本文所开展工作为大规模无人机群在协同态势感知、协同灾害救援、协同目标跟踪等任务场景中的应用提供了积极探索和支撑。

Large-scale UAV cooperative target coverage is an important research direction in the field of UAV cooperative decision-making, which has a wide range of application requirements and important theoretical research significance in both military and civilian fields. This dissertation takes large-scale unmanned aerial vehicles cooperatively detecting and covering ground targets in a dynamic environment as the background, aiming at the collaborative decision-making problem of UAV swarm maintaining communication connectivity and optimizing coverage of dynamic targets under the condition of only local information. Based on graph theory and game theory, the dissertation carries out systematic modeling, control architecture research, optimization algorithm design, and physical verification platform construction and verification. The main work and contributions include:

(1) Aiming at the problem of large-scale UAV cooperative target coverage, an in-depth analysis is made and a mathematical optimization model is established. At the same time, combined with relevant game theory knowledge, it is proved that large-scale UAV cooperative target coverage is an accurate potential game process. Therefore, there is a Nash equilibrium solution, which provides theoretical guidance for subsequent scheme design.

(2) A large-scale UAV cooperative target coverage method under a flat control architecture is proposed. This control architecture is mainly oriented to mission scenarios with small mission areas. First, in order to reduce the computational complexity of the algorithm, The task environment map. is gridded, and the target grouping mechanism is designed to simplify the target allocation process. Secondly, based on the theoretical guidance of the potential game above, a distributed local negotiation algorithm is designed in combination with the Consensus-based Auction Algorithm (CBAA) to achieve optimal coverage of ground targets while maintaining communication connectivity under distributed control. Finally, the validity of the method is tested by simulation test.

(3) A large-scale UAV cooperative target coverage method under a layered control architecture is proposed. This control architecture is mainly oriented to mission scenarios with large task area spans and a large number of targets. First of all, aiming at the low efficiency of resource scheduling in flat control, a top-level CPU is designed to be responsible for global resource management and scheduling. Secondly, in order to improve the efficiency of UAV collaborative decision-making, the targets are clustered and grouped. By proving that the large-scale UAV target coverage process is an exact potential game process, a recommended and select distributed algorithm is designed. The algorithm can make decisions without multiple rounds of negotiation iterations, and each UAV seeks a global suboptimal solution by maximizing its own utility function. Thirdly, in view of the huge amount of communication during the mission execution process of large-scale UAVs, which leads to communication congestion, a situation information understanding and extraction mechanism is designed, which makes the large-scale UAV system have better scalability and higher task execution efficiency. Finally, the validity of the method is tested by simulation test.

(4) A physical verification platform is designed and built, and the effectiveness of the above two target coverage methods is verified. Firstly, the network communication architecture and hardware system architecture design of the experimental platform are completed. Secondly, a semi-physical algorithm verification platform is built to conduct semi-physical experimental verification of the flat control architecture, and the effectiveness of the distributed negotiation algorithm is verified through multiple sets of experimental tests. Finally, based on the successful test of the semi-physical experiment, multiple physical UAVs/vehicles equipped with onboard computers are tested in kind, showing that the layered control method proposed has good target coverage performance in large-scale UAV cooperative target coverage and scalability.

To sum up, this dissertation systematically studies the large-scale UAV cooperative target coverage problem from the aspects of theoretical analysis, system control architecture, target optimal allocation and communication and maintenance, and carries out digital simulation, semi-physical and physical test verification. The work carried out provides active exploration and support for the application of large-scale UAV swarms in mission scenarios such as cooperative situation awareness, coordinated disaster rescue, and coordinated target tracking.