基于角度信息的单站无源定位跟踪算法研究

发布时间：2018-01-10 12:15

本文关键词：基于角度信息的单站无源定位跟踪算法研究　出处：《电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：利用角度信息对目标进行单站无源定位跟踪是非常重要的定位跟踪方法。该方法首先通过无源传感器阵列接收目标辐射的电磁波,从接收到的辐射波中提取出角度信息,再利用角度观测量与目标的状态向量之间的关系式来对目标进行定位跟踪,所以该方法对角度观测量的精度具有很高的要求。由于要分析目标和观测站的不同运对状态,因此对于系统的可观测性分析也是必不可少的。现介绍本文的主要内容:1.介绍了利用角度信息来对目标辐射源进行单站无源定位跟踪的原理,以及根据目标的不同运动形式,对系统的可观测性进行了分析。2.研究了基于子空间数据融合的多目标直接定位算法,传统的多目标定位过程一般需要多个步骤,首先对目标进行DOA(波达角)估计,再利用定位算法对目标的状态进行估计。而基于子空间数据融合的多目标直接定位算法则不需要进行DOA估计。观测站在不同的观测位置的对多个目标进行观测,得到一系列的观测数据,利用这些观测数据构造出以位置坐标为参数的代价函数,而且整个定位过程只通过对该代价函数进行一次最优估计,就可以估计出所有目标的位置信息,因此比传统的多目标定位算法更高效,精度更高。3.介绍了传统的利用角度信息对目标辐射源进行定位跟踪的滤波算法,简要介绍了跟踪模型,主要介绍了扩展卡尔曼滤波(EKF)算法,修正增益扩展卡尔曼滤波(MGEKF)算法和修正协方差扩展卡尔曼滤波(MVEKF)算法,以及不敏卡尔曼滤波算法(UKF),该算法只需要利用少量采样点进行非线性变换,不需要进行线性化处理,因此具有更高的估计精度。最后对这些滤波算法进行了仿真,并且分析了不同仿真场景下这些滤波算法的性能。4.研究了基于优化加权数据融合的UKF算法,利用单站多传感器来实现数据融合。将每个传感器的观测数据分别利用UKF算法进行滤波估计,得到局部的滤波估计值,再利用优化加权数据融合算法对局部的滤波估计值进行数据融合,得到融合后的滤波估计结果,通过这种处理滤波估计精度得到了提高。最后对算法进行了仿真实验和性能分析。
[Abstract]:It is very important to use angle information to track the target with single station passive location and tracking. This method first receives the electromagnetic wave from the target by passive sensor array. Angle information is extracted from the received radiation wave, and then the relationship between the angle observation and the state vector of the target is used to locate and track the target. Therefore, this method has a high requirement for the accuracy of the angular observations, because of the analysis of the target and the different operational pairs of the observation station. Therefore, it is necessary to analyze the observability of the system. This paper introduces the main content of this paper: 1. The principle of using angle information to track the target emitter is introduced. The observability of the system is analyzed according to the different moving forms of the target. 2. The multi-target direct location algorithm based on subspace data fusion is studied. The traditional multi-target localization process usually requires many steps. Firstly, the DOA (Angle of arrival) estimation of the target is carried out. Then the state of the target is estimated by using the localization algorithm. The multi-target direct localization algorithm based on subspace data fusion does not need to estimate DOA. The observation station at different observation positions is used to observe multiple targets. Test. A series of observation data are obtained, and the cost function based on the position coordinates is constructed by using these data, and the whole positioning process is based on only one optimal estimation of the cost function. The location information of all targets can be estimated, so it is more efficient and accurate than the traditional multi-target localization algorithm. The traditional filtering algorithm using angle information to locate and track the target emitter is introduced. The tracking model is briefly introduced, and the extended Kalman filter (EKF) algorithm is mainly introduced. Modified gain extended Kalman filter (MGEKF) algorithm, modified covariance extended Kalman filter (MVEKF) algorithm, and unsensitive Kalman filter algorithm (UKF). The algorithm only needs a small number of sampling points to carry out nonlinear transformation and does not need to be linearized, so it has higher estimation accuracy. Finally, these filtering algorithms are simulated. And the performance of these filtering algorithms in different simulation scenarios is analyzed. 4. The UKF algorithm based on optimized weighted data fusion is studied. The observation data of each sensor is estimated by UKF algorithm, and the local filter estimation value is obtained. Then the local filter estimation value is fused with the optimized weighted data fusion algorithm, and the fused filtering estimation results are obtained. The accuracy of the filter estimation is improved by this method. Finally, the simulation experiment and performance analysis of the algorithm are carried out.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN953

【相似文献】