摘要
针对预警监视和指挥决策领域对目标行为智能分析的迫切需求,从信息融合领域和数据挖掘领域的交叉点入手,围绕目标行为规律挖掘、目标异常行为检测、目标行为模式在线分类、目标异常行为实时检测等方面,对基于航迹数据挖掘的目标行为分析技术进行了概述分析,从理论框架、大数据的挑战、粒度控制、多因素制约、应用方向等方面,探讨了面临的困难与挑战,并提出了本领域具有研究价值的热点研究方向.
Aiming at the urgent demand of intelligent analysis of target behavior in the field of early warning surveillance and command decision,starting from the intersection of information fusion and data mining,this paper summarizes and analyzes the target behavior analysis technologies based on trajectory data mining from the aspects of target regular behavior mining,target anomaly behavior detection,target frequent behavior online classification,target anomaly behavior real-time detection.The difficulties and challenges are discussed from the theoretical framework,challenges of big data,granularity control,multiple factors restriction,application direction.Several hot research directions with research value in this field are put forward.
作者
潘新龙
程学旗
王海鹏
何友
李敏波
PAN Xin-Long;CHENG Xue-Qi;WANG Hai-Peng;HE You;LI Min-Bo(Institute of Information Fusion,Naval Aviation University,Yantai Shandong 264001,China;Key Laboratory of Network Data Science and Technology,Institute of Computing Technology,Chinese Academy of Sciences,Beijing 100090,China;Software School,Fudan University,Shanghai 200433,China)
出处
《指挥与控制学报》
CSCD
2021年第4期335-341,共7页
Journal of Command and Control
基金
国家自然科学基金(62076249,61671157)
山东省自然科学基金(ZR2020MF154)
山东省重点研发计划(2020CXGC010701,2020ZLYS11)
山东省高等学校青创科技支持计划(2019KJN031,2020KJN006)
上海科技创新行动计划(18511107800)资助。
关键词
目标航迹
数据挖掘
行为规律
异常行为
行为模式
target trajectory
data mining
regular behavior
anomaly behavior
frequent behavior
