抗“飞点”的UKF-GMPCPHD滤波算法  

New UKF-GMPCPHD algorithm for outliers' rejection

作　　者：黄伟平[1] 徐毓[1] 甘少武[2] 

机构地区：[1]空军雷达学院科研部,湖北武汉430019 [2]空军雷达学院训练部,湖北武汉430019

出　　处：《系统工程与电子技术》2012年第1期34-39,共6页Systems Engineering and Electronics

基　　金：国防预研基金资助课题

摘　　要：为实现被动测角目标状态和数目的实时估计,在高斯混合粒子(Gaussian mixture particle,GMP)的势化概率假设密度(cardinalized probability hypothesis density,CPHD)滤波框架下,提出了基于抗"飞点"无迹卡尔曼滤波器(unscented Kalman filter,UKF)的GMPCPHD滤波算法,即抗"飞点"的UKF-GMPCPHD滤波算法。在该算法中,粒子滤波的重要性采样函数由抗"飞点"UKF产生,粒子的预测与更新采用拟蒙特卡罗(quasi-MonteCarlo,QMC)方式,目标状态的概率假设密度(probability hypothesis density,PHD)和势分布用一组高斯粒子滤波器(Gaussian particle filtering,GPF)近似。通过该算法与GMPCPHD、UKF-GMPPHD滤波算法的对比仿真,验证了该算法良好的跟踪性能。To get the state and number estimation of the bearings-only targets in real time,employing the frame of a cardinalized probability hypothesis density（CPHD） for the Gaussian mixture particle（GMP）,a GMPCPHD algorithm based on unscented Kalman filter（UKF） for outliers＇ rejection,called UKF-GMPCPHD algorithm for outliers＇ rejection,is proposed.In the new algorithm,the sophisticated proposal distributions for the particle filter are generated by the UKF for outliers＇ rejection,and the prediction and update distributions for the particles are approached by quasi-Monte Carlo（QMC） method,and the probability hypothesis density（PHD） and cardinalized distributions are approximated by a mixture of Gaussian particle filtering（GPF）.Finally,the comparison has been done among the UKF-GMPCPHD,GMPCPHD and UKF-GMPPHD.Simulation results show the good tracking performance of the proposed algorithm.

关 键 词：非线性跟踪 目标数目 势化概率假设密度 门限函数 拟蒙特卡罗 

分 类 号：TN911.72[电子电信—通信与信息系统]