多维标度无源定位方法研究

发布时间：2018-02-10 16:17

本文关键词： 无源定位多维标度方法观测站位置误差观测站时钟偏差混合定位　出处：《电子科技大学》2017年博士论文　论文类型：学位论文

【摘要】：无源定位因具有探测距离远和隐蔽性强等优点而在国家安全与现代战争中发挥着越来越重要的作用。无源定位在广泛应用的同时,也面临着越来越多的挑战。本论文围绕观测站位置和时钟有误差的定位问题以及基于测距和测角的混合定位问题,对多维标度(MDS:Multidimensional Scaling)无源定位算法开展了研究。主要研究内容和取得的成果有:1.研究了观测站位置有误差的测距MDS定位算法。在测距MDS统一框架上提出了一种观测站位置有误差的测距MDS算法,这是一种闭式解法。理论分析表明,当测距误差与观测站位置误差服从零均值高斯分布时,该算法是一种无偏估计,当误差很小时能够达到定位的克拉美劳下界(CRLB:Cramer-Rao lower bound)。仿真结果验证了理论推导,并且表明当测距误差和观测站位置误差较大时,定位精度比两步加权最小二乘测距定位算法和加权MDS测距定位算法高。并且这三种算法的计算复杂度相当。2.研究了观测站有位置误差的时差MDS定位算法。提出了一种观测站位置有误差的时差MDS定位算法,这是一种闭式解法。理论分析表明,当时差测量误差与观测站位置误差服从零均值高斯分布时,该算法是一种无偏估计,在误差很小时能够达到CRLB。仿真结果验证了理论推导,同时也表明当测量误差和观测站位置误差较大时,算法的定位精度比两步加权最小二乘时差定位算法和加权MDS时差定位算法高。并且这三种算法的计算复杂度相当。3.研究了观测站位置有误差的时频差MDS定位算法。提出了一种观测站有位置误差的时频差MDS算法,这是一种闭式解。理论分析表明,当时频差测量误差与观测站位置误差服从零均值高斯分布时,算法对辐射源的位置和速度估计是无偏的,在误差很小时能够达到CRLB。仿真结果验证了理论推导,同时也表明当测量误差和位置误差较大时,算法定位精度比两步加权最小二乘时频差定位算法、约束总体最小二乘时频差定位算法和加权MDS时频差定位算法更高。并且提出的算法计算复杂度比约束总体最小二乘算法低,与其它两种算法相当。4.开展了观测站位置和时钟有误差的测距MDS多目标定位算法研究。推导了观测站位置和时钟有误差的测距定位的CRLB,通过CRLB分析了测距误差,观测站位置误差,观测站时钟误差对定位精度的影响,结果表明,通过观测站之间的距离测量可以降低观测站位置和时钟误差对定位精度的影响。提出一种观测站位置和时钟有误差的测距MDS多目标定位闭式解,通过仿真验证了算法的有效性。5.揭示了测距与测角混合MDS定位算法与子空间MDS算法的联系,它们是同一MDS内积矩阵的不同表达方式。提出了另外一种测距与测角混合MDS定位方法,这种方法只需一个观测站测角,降低了系统复杂度。通过仿真验证了当测角精度较高时,这种混合MDS定位算法比纯测距定位算法和只有一个站测角的Chan定位算法精度高。
[Abstract]:Passive location has long detection distance and strong concealment and other advantages in the national security and the modern war is playing an increasingly important role in the wide application of passive location. At the same time, is also facing more and more challenges. This paper focuses on the problem of positioning stations and the clock observation error and hybrid positioning and ranging angle based on multidimensional scaling (MDS:Multidimensional Scaling) passive location algorithm has been studied. The main research contents and achievements are: ranging MDS positioning algorithm 1. research station location error in the measured distance. The MDS unified framework proposes a MDS algorithm based on ranging station location error, this is a closed form solution. Theoretical analysis shows that, when the distance error and station location error obeys zero mean Gauss distribution, the algorithm is an unbiased estimator when the error is small enough to reach To lower positioning of the Cramer Rao lower (CRLB:Cramer-Rao lower bound). The simulation results verify the theoretical derivation, and show that when the distance error and station location error is large, the positioning accuracy of the two step weighted least square localization algorithm and weighted MDS location algorithm. And the calculation of the three algorithms in time MDS localization algorithm is quite complex.2. on the observation station with position error. The time difference MDS positioning algorithm for station location error, which is a closed form solution. Theoretical analysis shows that the time difference measurement error and observation position error obeys zero mean Gauss distribution, the algorithm is an unbiased estimator can achieve CRLB. simulation results verify the theoretical derivation in the error is very small, it also suggests that the position error of the measurement error and the observation station is larger than two, the positioning accuracy of the algorithm by weighted least step two TDOA location algorithm and weighted MDS algorithm and calculate the TDOA location. These three algorithms are complex frequency difference MDS positioning algorithm is quite.3. to study the station location error. The paper proposed a MDS algorithm with position error of observation station when the frequency difference, this is a closed solution of theoretical analysis. That was the frequency difference measurement error and station location error obeys zero mean Gauss distribution algorithm, the position and velocity of the radiation source estimation is unbiased, can achieve the CRLB. simulation results verify the theoretical derivation in the error is very small, but also shows that when the measurement error and the position error is large, the localization accuracy is more than two when the frequency difference step weighted least squares algorithm, difference frequency localization algorithm of differential frequency localization algorithm and weighted MDS constrained total least squares when higher. And the proposed algorithm computational complexity than the constrained total least squares algorithm is low, and The other two algorithms is carried out.4. location MDS stations and the clock observation error algorithm for multi target localization was deduced. The location of station position and clock observation error of CRLB, through the CRLB analysis of ranging error and position error of observation station, station clock errors, the accuracy of the results showed that the distance measurement can reduce the impact of the station location and clock error on the positioning accuracy of observation stations in between. Put forward a MDS ranging location of station and clock error location closed solution, through the simulation of.5. the effectiveness of the algorithm reveals the location and angle of hybrid MDS location algorithm and subspace MDS the algorithm, they are different expressions of the same MDS inner product matrix. It provides another angle ranging and hybrid MDS positioning method, this method only needs a low angle observation station, The complexity of the system is verified by simulation. When the angle measurement accuracy is high, the hybrid MDS localization algorithm has higher accuracy than the pure ranging location algorithm and the Chan location algorithm with only one station angle measurement.

【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TN95

【参考文献】