异构MAS下反导作战多传感器任务规划分层决策框架  

Research on layered decision-making of multi-sensor planning based on heterogeneous MAS in anti-TBM combat

作　　者：倪鹏[1,2] 刘进忙[1] 付强[1] 高嘉乐 

机构地区：[1]空军工程大学防空反导学院 [2]中国人民解放军94921部队

出　　处：《系统工程与电子技术》2016年第8期1816-1825,共10页Systems Engineering and Electronics

基　　金：国家自然科学青年基金(61102109)资助课题

摘　　要：综合利用多维传感器平台的资源互补优势进行协同探测和跟踪是反导作战的重要问题和难点。为提高传感器任务规划的效能,针对观测资源的异构性和任务的阶段性、动态性,引入任务共同体概念,剖析了反导作战多传感器任务规划问题的本质,形式化定了任务共同体下的行为准则和执行能力;在分析和建立的异构多Agent系统(multi-agent system,MAS)多传感器任务规划体系的基础上,深入探讨了集中式规划和分布式动态调整下的分层决策框架和求解方法,并通过仿真实验验证了所构建分层决策框架的有效性和合理性。It is an important and difficult problem for anti-TBM combat to realize detecting and tracking the TBMs cooperatively by comprehensively utilizing the complementary advantages of multi-dimensional sensor platforms.Aiming at the heterogeneous of sensor resources and the time-phased and dynamism of the mission,the community of interest（COI）is introduced to analyze the essence of multi-sensor mission planning（MSMP）.And then the behavior of COI is formally defined as well as its executive capability.On this basis,the architecture of MSMP based on heterogeneous multi-agent system（MAS）is analyzed and established.Finally,both centralized and distributed planning methods are also discussed in detail.Simulation results show the effectiveness and reasonability of the proposed method.

关 键 词：多传感器任务规划 任务共同体 异构多Agent系统 分层决策 双层规划 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]