有效去除定位偏差的TDOA/FDOA闭合解定位算法  

Effective bias reduction closed-form algorithm for source localization using TDOA and FDOA

作　　者：周成[1] 黄高明[1] 高俊[1] 

机构地区：[1]海军工程大学电子工程学院,湖北武汉430033

出　　处：《浙江大学学报（工学版）》2015年第12期2340-2348,2396,共10页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(60901069);国家"863"高技术研究发展计划资助项目(2013AAXXXX061)

摘　　要：针对现有的TDOA/FDOA无源定位算法存在定位偏差的问题,提出基于两步加权最小二乘估计器的偏差补偿算法.该算法具有闭合解的形式,通过对两步加权最小二乘算法在目标定位求解过程中的偏差值进行依次求解,得到最终的偏差值.通过对该偏差值进行数值仿真分析发现:偏差值随着噪声的增大变得越来越显著.采用将两步加权最小二乘算法的原始定位解减去其偏差值的方法,得到一个具有偏差补偿性质的定位结果.理论分析证明:在较低噪声的情况下,该算法对目标的定位结果能够达到克拉美罗界.将该算法与原两步加权最小二乘算法、Taylor算法和BiasRed算法进行比较,数值仿真结果表明:在噪声较低的情况下,该算法在保持目标定位的方差值与原算法一致的情况下,能够有效地减小目标的定位偏差,提高目标的定位性能.A bias reduction algorithm for well known two-step weighted least squares （WLS） estimator was proposed in order to reduce the source localization bias of the existing TDOA/FDOA algorithm. The algorithm has closed-form and the final bias of the source location was derived by deriving the location bias in each step of the two-step WLS estimator. The relationship of the noise and the bias was analyzed by simulations. Results show that the bias is found to become considerably large when noise level grows up Then, was de analysi Compa BiasRe the algorithm to reduce the bias was proposed. The source location estimate with compensated bias rived by subtracting the expected bias from the solution of the two-step WLS estimator. Theoretical s verifies that the accuracy of the estimator can attain CRLB accuracy when the noise is small. re the proposed method with the original two-step WLS estimator, the Taylor algorithm and the d algorithm through numerical simulation, results identify that the proposed method can keep the same mean square errors with the original algorithm and reduce the localization bias effectively when the noise is small.

关 键 词：被动定位 定位偏差 闭合解 到达时差(TDOA) 到达频差(FDOA) 两步加权最小二乘(WLS) 

分 类 号：TN97[电子电信—信号与信息处理]