基于多Agent的无人机群区域探测协同作战仿真

摘要

为帮助无人机群在区域探测协同作战中通过已有信息作出较为准确的判断从而有效选择探测路线,建立了基于多Agent的无人机群区域探测协同作战模型.依据所在战场环境、无人机群和其目标的特点等因素,提取出了无人机群的属性和行为,同时从体系结构视角设计了几种具有不同的信息共享方式和决策层级结构的模型,采用Repast建模仿真软件进行仿真试验.结果表明采取实时信息共享以及集中决策的方式往往能够得到准确度较高、冗余度可接受的探测结果.%A multi-agent-based model of area detection and cooperative combat for unmanned aerial vehicles (UAVs) cluster is designed for UAVs cluster to make a better judgement basing on existing information and effectively choose detect route in area detection and cooperative combat.Attributes and behaviors of UAVs are built according to characteristics of battlefield environment,UAVs and their targets.Models with different ways of information sharing and decision hierarchy are designed from the perspective of architecture and the concerned simulation experiments are carried out on Repast Simphony.The result shows that real-time information sharing and centralized decision achieves detection results with higher accuracy and acceptable redundancy.