Moment conditions for convergence of particle filters with unbounded importance weights

@article{Mbalawata2016MomentCF,
  title={Moment conditions for convergence of particle filters with unbounded importance weights},
  author={Isambi S. Mbalawata and Simo S{\"a}rkk{\"a}},
  journal={Signal Process.},
  year={2016},
  volume={118},
  pages={133-138}
}

Figures from this paper

Simo Sarkka Title : Year : Version : On the ( Non-) Convergence of Particle Filters with Gaussian Importance Distributions 2 01 5 Final
We consider convergence properties of particle filters with Gaussian importance distributions for certain time-varying Poisson regression models. We analyze both the classical
Resampling methods for particle filtering: identical distribution, a new method, and comparable study
TLDR
More than a dozen typical resampling methods are compared via simulations in terms of sample size variation, sampling variance, computing speed, and estimation accuracy, providing solid guidelines for either selection of existing resamplings methods or new implementations.
ON THE L P-CONVERGENCE OF A GIRSANOV THEOREM BASED PARTICLE FILTER
We analyze the Lp-convergence of a previously proposed Girsanov theorem based particle filter for discretely observed stochastic differential equation (SDE) models. We prove the convergence of the
On the LP-convergence of a Girsanov theorem based particle filter
  • S. Särkkä, É. Moulines
  • Mathematics
    2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
  • 2016
TLDR
The convergence of the algorithm with the number of particles tending to infinity is proved by requiring a moment condition and a step-wise initial condition boundedness for the stochastic exponential process giving the likelihood ratio of the SDEs.
Particles resampling scheme using regularized optimal transport for sequential Monte Carlo filters
A resampling method is presented for improving the performance of particle filters by using adaptive numbers of the resampling particles. The proposed method replaces the resampling step of a
Particle filtering with applications in networked systems: a survey
TLDR
This survey paper first provides an overview of the particle filtering methods as well as networked systems, and then investigates the recent progress in the design of particle filter for networked system.
Data Fusion With Inverse Covariance Intersection for Prior Covariance Estimation of the Particle Flow Filter
The prior covariance estimation method based on inverse covariance intersection (ICI) is proposed to apply the particle flow filter. The proposed method has better estimate performance and guarantees
A Survey of Recent Advances in Particle Filters and Remaining Challenges for Multitarget Tracking
TLDR
This review examines the intractable challenges raised within the general multitarget (multi-sensor) tracking due to random target birth and termination, false alarm, misdetection, measurement-to-track (M2T) uncertainty and track uncertainty.

References

SHOWING 1-10 OF 34 REFERENCES
Weight moment conditions for L4 convergence of particle filters for unbounded test functions
TLDR
Novel moment conditions for importance weights of sequential Monte Carlo based particle filters are derived, which ensure the L4 convergence of particle filter approximations of unbounded test functions.
Stability properties of some particle filters
Under multiplicative drift and other regularity conditions, it is established that the asymptotic variance associated with a particle filter approximation of the prediction filter is bounded
A Basic Convergence Result for Particle Filtering
TLDR
The basic nonlinear filtering problem for dynamical systems is considered, and a general framework, including many of the particle filter algorithms as special cases, is given.
A survey of convergence results on particle filtering methods for practitioners
TLDR
The aim of this paper is to present a survey of convergence results on particle filtering methods to make them accessible to practitioners.
Central limit theorem for sequential Monte Carlo methods and its application to Bayesian inference
The term sequential Monte Carlo methods or, equivalently, particle filters, refers to a general class of iterative algorithms that performs Monte Carlo approximations of a given sequence of
Optimality of the auxiliary particle filter
In this article we study asymptotic properties of weighted samples produced by the auxiliary particle filter (APF) proposed by Pitt and Shephard [17]. Besides establishing a central limit theorem
On the L4 convergence of particle filters with general importance distributions
  • I. Mbalawata, S. Särkkä
  • Engineering
    2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
  • 2014
TLDR
The result essentially shows that with general importance distributions the particle filter converges provided that the importance weights are bounded and by numerical simulations this condition is often also a practical requirement for a good performance of a particle filter.
A General Convergence Result for Particle Filtering
TLDR
The present contribution is that it is proved that the particle filter converge for unbounded functions in the sense of Lp-convergence, for an arbitrary p ≥ 2.
Convergence of Sequential Monte Carlo Methods
TLDR
This paper presents a general sequential Monte Carlo (SMC) method which includes most of the important features present in current SMC methods and generalizes and encompasses many recent algorithms.
...
...