作　　者： ; ;

机构地区： 中山大学新华学院信息科学系

出　　处： 《吉林师范大学学报（自然科学版）》 2011年第2期114-118,共5页

摘　　要： 提出一类改进的粒子滤波算法.对于建议分布的选取方案,此算法采取强跟踪分散的卡尔曼滤波方式建立它的建议分布.由于线性调节参数,此算法让系统拥有更优越的自适应性及鲁棒性,对高机动目标具有更强的跟踪效果,继而为强跟踪扩展卡尔曼滤波的能力.仿真结论说明,此算法的性能比别的几类非线性滤波算法更加优秀.比如辅助粒子滤波器(APF)、迭代扩展卡尔曼粒子滤波器(IEKF-PF)、Unscented粒子滤波器(UPF)、正则化粒子滤波器(RPF),则是在bootstrap粒子滤波器提出之后,继而出现的改进的粒子滤波器0基于粒子滤波,本文提出了阻止粒子退化的两个重点原因,以及选取合适的采样建议分布及重采样算法. This article puts forward one kind of particle filter algorithm.According to the selection of proposal distribution,this algorithm uses a method which is called The Strong tracking scattered Kalman filter to set up its proposal distribution.Because lines adjust the parameter,this system has the advantage of self-adaptability and robustness.To the high maneuvering target,it uses a stronger tracking effect to extend the performance of the Kalman filtering.The conclusion of simulation explains that the performance of algorithm is much more outstanding than other kinds of nonlinear filtering.For example,after the bootstrap Particle Filter was invented,auxiliary particle filter（APF）,Iterated extended kalman particle filter（IEKF-PF）、unscented particle filter（UPF）、regularization particle filter（RPF） appeared continuwhichley.Basing on the particle filter,this article mention two important reasons of preventing the particle degeneratively.In the meanwhile,it selects proposal distribution and resample technique properly to solve the problems.

关 键 词： 粒子滤波 密度函数 滤波算法 目标跟踪 卡尔曼滤波

领　　域： [自动化与计算机技术] [自动化与计算机技术]