基于UKF和优化组合策略的改进粒子滤波算法

发布时间：2018-01-26 16:04

  本文关键词： 粒子滤波 无迹卡尔曼滤波 优化组合策略 距离判决　出处：《计算机工程与科学》2017年08期 　论文类型：期刊论文

【摘要】：针对标准粒子滤波算法存在的粒子退化与贫化问题,提出了一种新的改进粒子滤波算法。该算法采用无迹卡尔曼滤波、优化组合策略和标准粒子滤波相结合的方法,运用UKF产生重要性密度函数,解决标准PF算法中以先验概率密度函数作为建议分布所引发的退化问题;运用优化组合重采样策略保证所有粒子的信息以一定概率得到继承,维持粒子集中粒子的多样性。理论分析与仿真结果均表明,改进算法能有效地解决标准粒子滤波存在的粒子退化问题并避免粒子贫化现象的出现,具有更高的状态估计精度。
[Abstract]:Aiming at the problem of particle degradation and dilution in standard particle filtering algorithm, a new improved particle filter algorithm is proposed, which uses unscented Kalman filter. Combining the optimal combination strategy and standard particle filter, using UKF to generate the importance density function, the degradation problem caused by the priori probability density function as the suggested distribution in the standard PF algorithm is solved. The optimal combinatorial resampling strategy is used to ensure that the information of all particles is inherited with a certain probability, and the diversity of particles in the particle concentration is maintained. The theoretical analysis and simulation results show that. The improved algorithm can effectively solve the particle degradation problem of standard particle filter and avoid the phenomenon of particle dilution. It has higher state estimation accuracy.
【作者单位】： 空军工程大学信息与导航学院;空军工程大学防空反导学院;
【分类号】：TN713
【正文快照】： 1引言在线性假设和高斯噪声背景条件下,卡尔曼滤波是雷达目标跟踪的最佳算法,而在现实世界中,目标模型大多是非线性非高斯的,常用的算法有扩展卡尔曼滤波EKF(Extended Kalman Filter)、无迹卡尔曼滤波UKF(Unscented Kalman Filter)以及粒子滤波PF(Particle Filter)。其中粒子，

本文编号：1466029