基于角度约束采样的单站无源定位混合粒子滤波算法
doi: 10.3724/SP.J.1146.2006.01340
Hybrid Particle Filtering Algorithm for Passive Location by a Single Observer Based on Bearing Constrained Sampling
-
摘要: 为实现固定单站对运动辐射源的快速定位，该文给出了一种基于角度约束采样的混合粒子滤波算法。该算法从EKF(Extended Kalman Filter)滤波得到建议分布，利用角度测量对状态变量的约束关系从建议分布产生所需粒子，可以减少粒子滤波用于高维情况时所需的粒子数目，改善滤波性能，降低运算成本。结合利用多普勒变化率和角度测量的单站定位方法，与EKF，UKF(Unscented Kalman Filter)以及一般混合粒子滤波算法的仿真比较表明，该算法在滤波收敛速度、跟踪精度以及稳定性方面优于其它算法，估计误差更接近Cramer-Rao下界。
-
关键词:
- 无源定位; 粒子滤波; 多普勒变化率
Abstract: To achieve fast location of moving emitter by a single stationary observer, an algorithm of hybrid particle filter based on bearing constrained sampling is presented. The algorithm gets proposal importance density from Extended Kalman Filter(EKF), and generates particles through the constraint between bearing measurements and the state variables, thus the number of particles and computation cost decrease when tackling high-dimensional filtering, and the filtering performance gets improved. Applying the algorithm to the location method of using Doppler changing rate and bearing measurements, simulation results of comparing the proposed algorithm with EKF, Unscented Kalman Filter(UKF) and the general hybrid particle filter, show that the proposed algorithm is superior in convergence speed, tracking precision and filtering stability to others, and the estimation error is more closer the Cramer-Rao lower bound. -
孙仲康，周一宇，何黎星. 单多基地有源无源定位技术. 北京：国防工业出版社，1996: 196-349.[2]Becker K. Three-dimension target motion analysis usingangle and frequency measurements[J].IEEE Trans. onAerospace and Electronic Systems.2005, 41(1):284-301[3]郭福成. 基于运动学原理的单站无源定位与跟踪关键技术研究. [博士论文]，国防科技大学，2002.[4]Julier S J and Uhlmann J K. Unscented filtering andnonlinear estimation[J].Proc. of The IEEE.2004, 92(3):401-422[5]Arulampalam M S, Maskell S, and Gordon N, et al.. Atutorial on particle filters for online nonlinear/non-GaussianBayesian tracking[J].IEEE Trans. on Signal Processing.2002,50(2):174-188[6]Freitas J F G. Bayesian method for neural networks. [Ph.D.dissertation], Cambridge University, 1999.[7]Liu J S and Chen R. Sequential Monte Carlo methods fordynamical systems. J. Amer. Statist. Assoc. 93, 1998: 1032-1044.[8]Bergman N. Recursive Bayesian estimation: Navigation andtracking applications. [Ph.D. dissertation], Link.pingUniversity, 1999. -
计量
- 文章访问数: 3155
- HTML全文浏览量: 90
- PDF下载量: 832
- 被引次数: 0
下载:
