基于多源数据融合的雷达跟踪状态判定  

Radar Tracking State Determination Based on Multi-source Data Fusion

作　　者：胡利坤 贾江磊 全刚 刘向锋 王凯 HU Likun;JIA Jianglei;QUAN Gang;LIU Xiangfeng;WANG Kai(Jiuquan Satellite Launch Center,Jiuquan,Gansu 732750)

机构地区：[1]中国酒泉卫星发射中心,甘肃酒泉732750

出　　处：《火控雷达技术》2025年第1期32-36,48,共6页Fire Control Radar Technology

摘　　要：针对当前航天测控中引导源数据来源单一、雷达跟踪状态判定准确率不高的问题,对多源测量数据进行分簇和融合处理,以提高引导数据的精度。首先,利用AGNES聚簇分类算法将多源测量数据分为一簇正常数据和多簇异常数据,通过剔除异常数据来提高融合数据的质量;其次,基于聚簇获取的正常数据,利用加权平均融合算法得到融合数据;最后,将融合数据应用到雷达跟踪状态判定。仿真结果表明,融合后的数据误差远小于原始测量数据,雷达跟踪状态判定结果准确率得到提高。In order to solve the problems of single source of guided source data and low accuracy of radar tracking state determination in the current test task,the multi-source measurement data are clustered and fused,to improve the accuracy and stability of data.Firstly,AGNES clustering algorithm was used to divide multi-source measurement data into one cluster of normal data and multi-cluster of abnormal data,which can improve the quality of fusion data by eliminating abnormal data,the weighted average fusion algorithm is used to get the fusion data,and finally,the fusion data was applied to determine the radar tracking state.The simulation results showed that the data error after fusion was much less than the original measurement data,and the accuracy of radar tracking state judgment was improved.

关 键 词：雷达 数据融合 AGNES聚簇 跟踪状态判定 

分 类 号：TN95[电子电信—信号与信息处理]