王浩淼, 袁莞迈, 马可, 张颖, 申燕凯, 魏晨. 基于联盟博弈的无人机/无人车异构集群验证[J]. 工程科学学报, 2024, 46(7): 1207-1215. DOI: 10.13374/j.issn2095-9389.2023.10.25.002
引用本文: 王浩淼, 袁莞迈, 马可, 张颖, 申燕凯, 魏晨. 基于联盟博弈的无人机/无人车异构集群验证[J]. 工程科学学报, 2024, 46(7): 1207-1215. DOI: 10.13374/j.issn2095-9389.2023.10.25.002
WANG Haomiao, YUAN Wanmai, MA Ke, ZHANG Ying, SHEN Yankai, WEI Chen. Verification of a UAV/UGV heterogeneous swarm based on coalitional game theory[J]. Chinese Journal of Engineering, 2024, 46(7): 1207-1215. DOI: 10.13374/j.issn2095-9389.2023.10.25.002
Citation: WANG Haomiao, YUAN Wanmai, MA Ke, ZHANG Ying, SHEN Yankai, WEI Chen. Verification of a UAV/UGV heterogeneous swarm based on coalitional game theory[J]. Chinese Journal of Engineering, 2024, 46(7): 1207-1215. DOI: 10.13374/j.issn2095-9389.2023.10.25.002

基于联盟博弈的无人机/无人车异构集群验证

Verification of a UAV/UGV heterogeneous swarm based on coalitional game theory

  • 摘要: 无人异构集群相较于单一类型、单一个体的无人平台,能够完成更为复杂的任务,同时对严苛战场环境有着更高的适应度. 在无人异构集群协同执行任务时,任务分配是至关重要的环节,需要考虑异构无人平台和任务的多种约束和目标. 传统的任务分配方法分配效率低且难以处理大规模复杂任务. 联盟博弈通过形成由若干参与者组成的联盟,根据个体的属性、偏好对群体进行划分,从而实现个体以及群体利益的最大化. 本文以无人异构集群任务分配为背景,研究了基于改进联盟博弈算法的最优分配策略,基于可能的战场环境设计了模拟任务场景并完成实验验证. 首先,考虑异构平台在任务中的初始位置、速度、携带资源以及个体声誉等因素,建立了基于空间自适应博弈(Spatial adaptive play algorithm, SAP)的联盟博弈的任务分配算法模型. 其次,基于任务场景,搭建了任务所需的软件与硬件平台. 最后,针对模拟的战场环境,对所提算法及搭建的异构无人集群平台进行了实验验证. 验证结果表明,在异构无人集群平台重分配的任务背景下,本平台能综合考虑战场态势,寻找最优的任务分配方式,协调各作战单位完成任务目标.

     

    Abstract: Unlike unmanned platforms comprising a single type or individual, heterogeneous unmanned swarm platforms excel in performing more intricate tasks and exhibit heightened adaptability to challenging battlefield environments. When coordinating tasks on a heterogeneous unmanned swarm platform, task assignment is a crucial aspect that requires considering various constraints and objectives related to the heterogeneous unmanned platforms and the tasks. Traditional task assignment methods exhibit low efficiency and handle large-scale complex tasks with difficulty. The alliance game divides a group according to the attributes and preferences of the individual by forming an alliance composed of several participants to maximize the interests of the individual and the group. This paper focuses on investigating optimal task assignment strategies within unmanned heterogeneous swarm platforms, employing an enhanced coalitional game algorithm. First, by considering the initial position, speed, resources, and individual reputation of heterogeneous platforms in the task, a task allocation algorithm is proposed on the basis of a coalitional game with a spatial adaptive play (SAP) mechanism. The SAP algorithm is an adaptive learning method employed in spatial games, capable of self-adjustment based on the specific characteristics of task allocation problems. Notably, this algorithm tends to randomly select and update target individuals during each iteration process with equal probability. However, a challenge arises when attempting random agent selection in a distributed environment. To address this issue, a periodic adaptive selection mechanism is introduced to facilitate periodic updating. Furthermore, to enhance the convergence speed of SAP, neighbor information and historical information are considered during task updates. Second, in alignment with the task scenario, the necessary software and hardware platform for task execution is established. Finally, the proposed algorithm and the unmanned heterogeneous swarm platform are verified experimentally on the basis of the simulated battlefield environment. The mission objectives of each unit are reassigned on the basis of the environmental situation and resource integration of each combat unit. At the same time, when the battlefield situation changes, such as with the emergence of new targets or the elimination of the current target, the platform will allocate new combat units to form new alliances with the existing units. This allocation ensures the optimal use of resources and guarantees the completion of missions. The verification results demonstrate that the platform possesses comprehensive situational awareness, enabling it to respond promptly to battlefield changes, make informed judgments, optimize task allocation, and effectively coordinate combat units to achieve mission objectives.

     

/

返回文章
返回