摘要
在多无人机协同突防过程中,针对无人机需完成目标分配与探干侦动作选择的任务决策存在模型构建难,对模型求解方法复杂度高的问题,提出一种蒙特卡罗树搜索(Monte Carlo tree search,MCTS)的改进方法实现多无人机多任务联合决策。首先,考虑无人机与雷达对抗中的角度、距离等态势要素,以及当前无人机动作执行成功概率和雷达状态有效概率,构建多无人机目标分配与探干侦动作统一决策数学模型。其次,提出搜索次数自适应调整的改进MCTS算法对模型求解,实现大规模解空间在线快速寻优。仿真结果表明,改进算法使多雷达系统对无人机的威胁程度下降约16.8%,相比多臂赌博机算法效果提升约5.08%,决策时间约0.23 s,比传统MCTS缩短约45.7%,有助于提升无人机战场生存率。
In the process of multiple unmanned aerial vehicles(UAV)cooperative penetration,to address the challenges of model construction and the high complexity of solving target allocation and detection jamming and reconnaissance action selection tasks,an improved method of Monte Carlo tree search(MCTS)is proposed to realize joint multi-mission decision making for multiple UAVs.First,a unified mathematical decision model for multi-UAV target allocation and reconnaissance actions is constructed,considering the situational factors such as angle and distance in the confrontation between UAV and radars,as well as the current probability of successful execution of UAV actions and the effective probability of radar states.An improved MCTS algorithm with adaptive adjustment of search frequency is proposed for fast online optimization in large solution spaces.Simulation results show that the improved algorithm decreases the threat level of multi-radar system to UAVs by about 16.8%,improves the effect of the algorithm compared to the multi-armed gambling machine by about 5.08%,and the decision time is about 0.23 s,which is about 45.7%shorter than that of the traditional MCTS,thereby improving UAV battlefield survivability.
作者
魏建林
林彦超
唐慧龙
张旺
王伟
WEI Jianlin;LIN Yanchao;TANG Huilong;ZHANG Wang;WANG Wei(College of Intelligent Systems Science and Engineering,Harbin Engineering University,Harbin 150001,China;Beijing Aerospace Automatic Control Institute,Beijing 100854,China)
出处
《系统工程与电子技术》
北大核心
2026年第2期556-568,共13页
Systems Engineering and Electronics
基金
国家自然科学基金(62271163)资助课题。
