综合运用模式识别及回归分析的数据融合算法.pdfVIP

DOI ：10.13195/j.cd.2010.01.113.huangwp.003 第25 卷第1 期 控 制 与 决 策 2010 年1 月 Vol. 25 No . 1 Cont ro l and D ec is io n Jan. 2010 : 1001-0920(2010) 01-0110-05 1a 1b 2 黄伟平 , 徐 毓 , 王 杰 ( 1. a. , b. , 430019; 2. 95174 , 430019) : 在混合式系统融合框架下, 提出综合运用模式识别和回归分析的数据融合 法. 在融合中心, 首先处理的是 各目标航迹起始段的本地航迹, 其次是经粗处理的原始测量. 基于模式相似性测度分析形成起始段的融合航迹, 并利 用融合航迹信息对原始测量数据进行回归分析, 估计融合点和置信区间, 跟踪滤波形成航迹. 通过仿真实验, 验证了 法的有效性和实用性. : 本地航迹; 初始段; 观测数据; 相似性测度; 回归分析; 数据融合 : TN911. 72 : A Algorithm of data fusion based on mode recognition and regression analysis H UA N G Wei-p ing 1a, X U Yu 1b , WA N G J ie 2 ( 1a. Gro p of Grad ate Management, 1b. Department for Scientific Research, Air Force Rader Academy, W han 430019, China; 2. Unit 95174 of PLA, W han 430019, China. Correspondent: HUANG Wei-ping, E-mail: kjld_ hw p@ 163. com) Abstract: An algorithm of data f sion is presented based on mode recognition and regression analysis in term s of mixed f sion system. Firstly , the initial tracks of every sensor on its own are sent to the centre, and the original observation data is obtained. To get the f sion track, the association algorithm based on pat tern similarity analysis is sed. When the sef l information of this short track is picked- p, a regression analysis is given to the original observation data w hich has a simple disposal. The f se point estimate and interval estimates of the meas red q antity can be obtained, and the object is tracked. Sim lation res lts show the effectiveness and feasibility of the method. Key words: Local track; Initial period; Observation data; Data similarity analysis; Regression analysis; Data f sion 1 引 言 . , ,