基于二阶中心差分滤波的高斯混合粒子PHD多目标跟踪算法  

Gaussian Mixture Particle PHD Filter Based on Second-Order Central Difference Filtering for Multi-target Tracking

作　　者：冉星浩 陶建锋[1] 贺思三[1] RAN Xinghao;TAO Jianfeng;HE Sisan(Air and Missile Defense College,Air Force Engineering University,Xi an 710051,China)

机构地区：[1]空军工程大学防空反导学院,陕西西安710051

出　　处：《探测与控制学报》2018年第6期68-73,共6页Journal of Detection & Control

基　　金：国家自然科学基金项目资助(61501495)

摘　　要：针对杂波环境下高斯混合粒子PHD面临的跟踪精度低、滤波发散等问题,提出基于二阶中心差分滤波的高斯混合粒子PHD算法。该算法在传统的高斯混合粒子PHD滤波的基础上,采取二阶中心差分滤波方法来得到最优的重要性密度函数,充分利用了量测信息对采样粒子进行更新,使得粒子分布更加接近目标真实的后验分布,然后对PHD进行更新。仿真结果表明,该算法能在杂波环境下有效地跟踪多个目标,与高斯混合粒子PHD算法相比,状态估计更加接近真实值,大大提高了跟踪精度和系统稳定性。In order to overcome the lower estimating accuracy and filtering divergence of traditional GMP-PHD algorithm in clutter environment,a new improved GMP-PHD was proposed in this paper,which based on the second-order central difference filter.Compared with the traditional GMP-PHD,second-order central difference filter was used to generate the importance density function,and the latest measurements were fully utilized,which enabled the distribution of particles to be much closer to the real posterior distribution of the target,and then to update a PHD.The theory analysis and simulation result showed that the improved algorithm could effectively track multiple targets in clutter environment,compared with GMP-PHD filter,and the result was closer to the true value.

关 键 词：概率假设密度滤波 二阶中心差分滤波 多目标跟踪 高斯混合粒子PHD 

分 类 号：TN953[电子电信—信号与信息处理]