西安工业大学学报
西安工業大學學報
서안공업대학학보
JOURNAL OF XI'AN TECHNOLOGICAL UNIVERSITY
2015年
1期
23-28
,共6页
徐振华%郑恒%李江源%刘光宏
徐振華%鄭恆%李江源%劉光宏
서진화%정항%리강원%류광굉
检测融合%总功率约束%偏移系数极大化%最大特征值问题
檢測融閤%總功率約束%偏移繫數極大化%最大特徵值問題
검측융합%총공솔약속%편이계수겁대화%최대특정치문제
detection fusion%total power constraint%deflection coefficient maximization%largest eigenvalue problem
针对传感器融合系统总发射功率约束条件下的相干检测融合问题,利用偏移系数极大化理论优化系统检测性能。建立了基于多址接入信道传输的多传感器分布式相干检测融合模型,根据融合中心对信道增益信息的获取程度,提出了信道增益信息完美已知时的总发射功率约束相干检测融合方法(T PC-DCM 方法)和信道增益统计信息已知时的总发射功率约束相干检测融合方法(T PC-DCM-CS方法),上述两种方法分别将总发射功率约束分布式检测融合问题转化为一个秩为1的最大特征值问题,获得了最优加权乘性系数的闭合解。蒙特卡洛仿真试验表明:在低虚警概率条件下,当信噪比大于等于0 dB时,文中提出的方法的性能优于传统的基于并行网络拓扑结构的最优似然比方法。
針對傳感器融閤繫統總髮射功率約束條件下的相榦檢測融閤問題,利用偏移繫數極大化理論優化繫統檢測性能。建立瞭基于多阯接入信道傳輸的多傳感器分佈式相榦檢測融閤模型,根據融閤中心對信道增益信息的穫取程度,提齣瞭信道增益信息完美已知時的總髮射功率約束相榦檢測融閤方法(T PC-DCM 方法)和信道增益統計信息已知時的總髮射功率約束相榦檢測融閤方法(T PC-DCM-CS方法),上述兩種方法分彆將總髮射功率約束分佈式檢測融閤問題轉化為一箇秩為1的最大特徵值問題,穫得瞭最優加權乘性繫數的閉閤解。矇特卡洛倣真試驗錶明:在低虛警概率條件下,噹信譟比大于等于0 dB時,文中提齣的方法的性能優于傳統的基于併行網絡拓撲結構的最優似然比方法。
침대전감기융합계통총발사공솔약속조건하적상간검측융합문제,이용편이계수겁대화이론우화계통검측성능。건립료기우다지접입신도전수적다전감기분포식상간검측융합모형,근거융합중심대신도증익신식적획취정도,제출료신도증익신식완미이지시적총발사공솔약속상간검측융합방법(T PC-DCM 방법)화신도증익통계신식이지시적총발사공솔약속상간검측융합방법(T PC-DCM-CS방법),상술량충방법분별장총발사공솔약속분포식검측융합문제전화위일개질위1적최대특정치문제,획득료최우가권승성계수적폐합해。몽특잡락방진시험표명:재저허경개솔조건하,당신조비대우등우0 dB시,문중제출적방법적성능우우전통적기우병행망락탁복결구적최우사연비방법。
This paper is mainly about the total power constrained distributed coherent detection .The deflection coefficient maximization (DCM ) is used to optimize the performance of the fusion system under the total power constraint .The model of the distributed coherent detection fusion system based on the multiple access channel (MAC) was established .According to the information access degree of the fusion center ,in each scenario ,two cases were studied ,including the other with the perfect channel gain information and the other with the statistical channel gain information .Two methods named TPC-DCM and TPC-DCM-CS proposed .The problem of distributed detection fusion under total power constraints of the system was transformed into the largest eigenvalue problem with a rank one by the two methods with the closed-form solutions of the optimal weighting coefficients obtained .Monte-Carlo simulations indicate that ,when the SNR is greater than or equal to 0 dB and the false alarm probability is low ,the proposed method is superior in the detection performence to the existing LRT method which is the optimal based on the parallel access channel (PAC) scheme .