摘要
针对态势认知中目标数量多、信息不确定、数据不精确等问题,提出一种基于区间数聚类的目标分群算法。首先,考虑到传感器测量数据具有误差且数据不完全等因素,采用区间数对传感器探测到的目标进行特征描述。然后,为有效利用区间数信息定义了一种新的距离度量,并给出了改进的区间数聚类目标分群算法。最后,构造4类相互独立的区间数据集,对区间数据进行分类测试,并通过典型想定场景设定多类目标实体,基于目标空间位置、运动特征和属性等要素进行空间分群和任务分群。仿真结果验证了算法能够有效对目标进行分群,具有较强的稳定性。
In order to solve the problems of large number of targets and uncertain and inaccurate information in situation,a target grouping algorithm based on interval number clustering is proposed.Firstly,aiming at the problem of sensor measurement data with error and incomplete,the interval number is used to describe the characteristics of the target detected by the sensor.Then,a new distance measure is defined to make use of interval number information effectively,and an improved interval number clustering target grouping algorithm is given.Finally,four types of independent interval data sets are constructed to classify and test the interval data.Through typical scenario,multiple types of target entities are set,and spatial groupring and task grouping are carried out based on the elements such as target spatial location,motion characteristics and attributes.Simulation results show that the algorithm can effectively group targets and has strong stability.
作者
王海滨
关欣
衣晓
WANG Haibin;GUAN Xin;YI Xiao(Naval Aviation University, Yantai 264001, China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2022年第2期577-583,共7页
Systems Engineering and Electronics
基金
国防科技卓越青年人才基金(2017-JCJQ-ZQ-003)
泰山学者工程专项经费(ts201712072)资助课题。
关键词
目标分群
区间数
聚类
态势认知
target grouping
interval number
clustering
situational cognitive
