基于差分进化改进粒子滤波的多径估计算法

发布时间：2018-04-14 21:28

本文选题：状态估计 + 粒子滤波　；参考：《太原理工大学学报》2017年01期

【摘要】：多径干扰因具有位置上的不相关性、不确定性等特点,不能通过差分技术来消除,成为高精度定位的主要误差源之一。因此,估计多径参数对抑制多径误差、提高导航系统的定位精度具有重要意义。本文将多径估计问题转化为状态空间模型下的参数估计问题,并利用粒子滤波(PF)进行多径估计。同时,为了克服标准PF存在粒子枯竭、导致估计结果可能收敛到错误值的问题,提出了基于差分进化改进粒子滤波(DEPF)的多径估计算法,该算法利用差分进化(DE)算法代替PF的重采样来产生新粒子,使新粒子朝着状态真实后验概率密度分布的方向移动,避免了重采样后粒子可能收敛到局部最优值的问题。仿真结果表明,在非高斯噪声下与基于PF和EKF的多径估计算法相比,本文算法具有更好的多径估计性能。
[Abstract]:Multipath interference is one of the main error sources in high precision positioning because of its irrelevance and uncertainty, which can not be eliminated by differential technique.Therefore, it is important to estimate multipath parameters to suppress multipath error and improve positioning accuracy of navigation system.In this paper, the multipath estimation problem is transformed into the parameter estimation problem under the state space model, and the particle filter (PFF) is used to estimate the multipath.At the same time, in order to overcome the problem that particle depletion exists in standard PF and the estimation result may converge to the wrong value, a multipath estimation algorithm based on improved particle filter (DEPF) based on differential evolution is proposed.In this algorithm, the new particle is generated by replacing the resampling of PF with the differential evolution (DED) algorithm, which makes the new particle move towards the direction of the real posteriori probability density distribution of the state, and avoids the problem that the particle may converge to the local optimal value after resampling.The simulation results show that the proposed algorithm has better multipath estimation performance than that based on PF and EKF under non- noise.
【作者单位】：太原理工大学信息工程学院;
【基金】：国家自然科学基金项目资助:基于统计信息集的非高斯系统多目标优化控制及性能评估策略研究(61503271,61603267) 山西省自然科学基金项目资助(20140210022-7)
【分类号】：TN713

【相似文献】