利用MKLD准则的自适应PF算法设计及其应用
|
摘要
为降低PF算法的计算量,提出了基于最大Kullback-Leibler距离(MKLD)准则的PF-AMCMC算法。该算法可在自适应地选择粒子数的前提下,同时自适应地选择粒子滤波算法中MCMC移动步骤实施的时刻,在保证一定的状态估计精度的条件下,减少粒子滤波的计算量。大量的数值试验和GPS/DR组合导航仿真试验表明,本文提出的算法较标准粒子滤波算法在克服粒子滤波计算量大的缺陷方面有显著的效果,且获得了精度更高的状态估计。
This paper presents a new algorithm named PF-AMCMC based on Maximum KullbackLeibler distance(abbreviated as MKLD)criterion.This algorithm can adaptively choose the number of particles and at the same time select the implementation moment of MCMC movement,and reduces the computational complexity under conditions guaranteeing the accuracy of state estimation.The results of computational experiments and a GPS/DR integrated navigation simulation experiment show that the improved particle filtering methods proposed in this paper have a better performances for state estimation than other approaches.
This paper presents a new algorithm named PF-AMCMC based on Maximum KullbackLeibler distance(abbreviated as MKLD)criterion.This algorithm can adaptively choose the number of particles and at the same time select the implementation moment of MCMC movement,and reduces the computational complexity under conditions guaranteeing the accuracy of state estimation.The results of computational experiments and a GPS/DR integrated navigation simulation experiment show that the improved particle filtering methods proposed in this paper have a better performances for state estimation than other approaches.
引文
[1]Koller D,Fratkina R.Using Learning for Approximation in Stochastic Processes[C].ICML-98,Wisconsin,1998
[2]Fox D.Adapting the Sample Size in Particle Filters Through KLD-sampling[J].International Journal of Robotics Research,2003,22(12):985-1 003
[3]Cui Pingyuan,Zheng Lifang,Pei Fujun.Researches on Integrated Navigation Methods Based on Adjusted Particle Filtering[J].Computer Engineering,2008a,34(14):185-187(崔平远,郑黎方,裴福俊.基于自调整粒子滤波的组合导航方法研究[J].计算机工程,2008a,34(14):185-187)
[4]Cheng Shuiying,Zhang Jianyun.Overview on Particle Filtering[J].Journal of Austronautics,2008,29(4):1 099-1 111(程水英,张剑云.粒子滤波评述[J].宇航学报,2008,29(4):1 099-1 111)
[5]Cui Pingyuan,Sun Xinrui,Pei Fujun.Researches on Strapdown Initial Alignment Method Based on Particle Filtering[J].Journal of System Simulation,2008b,20(20):5 714-5 717,5 721(崔平远,孙新蕊,裴福俊.一种基于自适应粒子滤波的捷联初始对准方法研究[J].系统仿真学报,2008b,20(20):5 714-5 717,5 721)
[6]Legland F,Oudjane N.A Sequential Particle Algorithm that Keeps the Particle System Alive[DB/OL].Rapport de Recherche 5826,INRIA.URL ftp://ftp.inria.fr/INRIA/publication/publi-pdf/RR/RR-5826.pdf,2006
[7]Yu Xingwei,Wang Shouyong.An Approach on Particle Filtering Based on Choices of Important Weight Value[J].Journal of Air Force Radar Academy,2009,23(1):1-3,6(于兴伟,王首勇.一种基于重要性权值选择的粒子滤波方法[J].空军雷达学院学报,2009,23(1):1-3,6)
[8]Cai Zelin,Li Kaican.Maximum Kullback-Leibler Distance of Usual Distribution[J].Journal of Wuhan University(Natural Science Edition),2007,53(5):513-517(蔡择林,李开灿.常见分布的最大Kullback-Leibler距离[J].武汉大学学报(理学版),2007,53(5):513-517)
[9]Merwe R V D.Sigma-point Kalman Filter for Probabilistic Inference in Dynamic State-space Models[D].Oregon:Oregon Health&Science University,2004
[10]Andrede Netto M L,Gimeno L,Mendes M J.On the Optimal and Suboptimal Nonlinear Filtering Problem for Discrete-time Systems[J].IEEE Transactions on Automatic Control,1978,23:1 062-1 067
[11]Fu Mengyin,Deng Zhihong,Zhang Jiwei.Theories on Kalman Filtering and Its Applicationsin Navigation System[M].Beijing:Science Press,2003(付梦印,邓志红,张继伟.Kalmanl滤波理论及其在导航系统中的应用[M].北京:科学出版社,2003)
[2]Fox D.Adapting the Sample Size in Particle Filters Through KLD-sampling[J].International Journal of Robotics Research,2003,22(12):985-1 003
[3]Cui Pingyuan,Zheng Lifang,Pei Fujun.Researches on Integrated Navigation Methods Based on Adjusted Particle Filtering[J].Computer Engineering,2008a,34(14):185-187(崔平远,郑黎方,裴福俊.基于自调整粒子滤波的组合导航方法研究[J].计算机工程,2008a,34(14):185-187)
[4]Cheng Shuiying,Zhang Jianyun.Overview on Particle Filtering[J].Journal of Austronautics,2008,29(4):1 099-1 111(程水英,张剑云.粒子滤波评述[J].宇航学报,2008,29(4):1 099-1 111)
[5]Cui Pingyuan,Sun Xinrui,Pei Fujun.Researches on Strapdown Initial Alignment Method Based on Particle Filtering[J].Journal of System Simulation,2008b,20(20):5 714-5 717,5 721(崔平远,孙新蕊,裴福俊.一种基于自适应粒子滤波的捷联初始对准方法研究[J].系统仿真学报,2008b,20(20):5 714-5 717,5 721)
[6]Legland F,Oudjane N.A Sequential Particle Algorithm that Keeps the Particle System Alive[DB/OL].Rapport de Recherche 5826,INRIA.URL ftp://ftp.inria.fr/INRIA/publication/publi-pdf/RR/RR-5826.pdf,2006
[7]Yu Xingwei,Wang Shouyong.An Approach on Particle Filtering Based on Choices of Important Weight Value[J].Journal of Air Force Radar Academy,2009,23(1):1-3,6(于兴伟,王首勇.一种基于重要性权值选择的粒子滤波方法[J].空军雷达学院学报,2009,23(1):1-3,6)
[8]Cai Zelin,Li Kaican.Maximum Kullback-Leibler Distance of Usual Distribution[J].Journal of Wuhan University(Natural Science Edition),2007,53(5):513-517(蔡择林,李开灿.常见分布的最大Kullback-Leibler距离[J].武汉大学学报(理学版),2007,53(5):513-517)
[9]Merwe R V D.Sigma-point Kalman Filter for Probabilistic Inference in Dynamic State-space Models[D].Oregon:Oregon Health&Science University,2004
[10]Andrede Netto M L,Gimeno L,Mendes M J.On the Optimal and Suboptimal Nonlinear Filtering Problem for Discrete-time Systems[J].IEEE Transactions on Automatic Control,1978,23:1 062-1 067
[11]Fu Mengyin,Deng Zhihong,Zhang Jiwei.Theories on Kalman Filtering and Its Applicationsin Navigation System[M].Beijing:Science Press,2003(付梦印,邓志红,张继伟.Kalmanl滤波理论及其在导航系统中的应用[M].北京:科学出版社,2003)