Analysis Of Error Propagation In Particle Filters With Approximation
Help pages Full-text links: Download: PDF PostScript Other formats (license) Current browse context: math.PR
Download Full-text PDF Analysis of error propagation in particle filters with approximationArticle (PDF Available) in The Annals of Applied Probability 21(2011) · August 2009 with 22 ReadsDOI: 10.1214/11-AAP760 · Source: arXiv1st Boris Oreshkin18.67 · McGill University2nd Mark J. CoatesAbstractThis paper examines the impact of approximation steps that become necessary when particle filters are implemented on resource-constrained platforms. We consider particle filters that perform intermittent approximation, either by subsampling the particles or by generating a parametric approximation. For such algorithms, we derive time-uniform bounds on the weak-sense $L_p$ http://arxiv.org/abs/0908.2926 error and present associated exponential inequalities. We motivate the theoretical analysis by considering the leader node particle filter and present numerical experiments exploring its performance and the relationship to the error bounds.Discover the world's research10+ million members100+ million publications100k+ research projectsJoin for free arXiv:0908.2926v1 [math.PR] 20 Aug 2009Submitted to the Annals of Applied ProbabilityarXiv: math.PR/0000000ANALYSIS OF ERROR PROPAGATION https://www.researchgate.net/publication/45868278_Analysis_of_error_propagation_in_particle_filters_with_approximation IN PARTICLEFILTERS WITH APPROXIMATION∗By Boris N. Oreshkin and Mark J. CoatesMcGill UniversityThis paper examines the impact of approximation steps that be-come necessary when particle filters are implemented on resource-constrained platforms. We consider particle filters that perform in-termittent approximation, either by subsampling the particles or bygenerating a parametric approximation. For such algorithms, we de-rive time-uniform bounds on the weak-sense Lperror and presentassociated exponential inequalities. We motivate the theoretical anal-ysis by considering the leader node particle filter and present numeri-cal experiments exploring its performance and the relationship to theerror bounds.1. Introduction. Particle filters have proven to be an effective ap-proach for addressing difficult tracking problems [9]. Since they are morecomputationally demanding and require more memory than most other fil-tering algorithms, they are really only a valid choice for challenging prob -lems, for which other well-established techniques perform poorly. Such prob-lems involve (approximated) dynamics and/or observation models that aresubstantially non-linear and non-Gaussian.A particle filter maintains a set of “particles” that are candidate statevalues of the system (for example, the position and velocity of
be down. Please try the request again. Your cache administrator is webmaster. Generated Fri, 30 Sep 2016 11:48:47 GMT by s_hv1002 (squid/3.5.20)