Uniform observability of hidden Markov models and filter stability for unstable signals

DOI10.1214/08-AAP576zbMATH Open1165.93034arXiv0804.2885MaRDI QIDQ2389607FDOQ2389607

Authors: Ramon van Handel 

Publication date: 17 July 2009

Published in: The Annals of Applied Probability (Search for Journal in Brave)

Abstract: A hidden Markov model is called observable if distinct initial laws give rise to distinct laws of the observation process. Observability implies stability of the nonlinear filter when the signal process is tight, but this need not be the case when the signal process is unstable. This paper introduces a stronger notion of uniform observability which guarantees stability of the nonlinear filter in the absence of stability assumptions on the signal. By developing certain uniform approximation properties of convolution operators, we subsequently demonstrate that the uniform observability condition is satisfied for various classes of filtering models with white-noise type observations. This includes the case of observable linear Gaussian filtering models, so that standard results on stability of the Kalman--Bucy filter are obtained as a special case.

Full work available at URL: https://arxiv.org/abs/0804.2885

Recommendations

• Observability and nonlinear filtering
• Discrete time nonlinear filters with informative observations are stable
• Stability properties of some particle filters
• Stability of nonlinear filters in nonmixing case
• Stability and uniform approximation of nonlinear filters using the Hilbert metric and application to particle filters

zbMATH Keywords

predictionhidden Markov modelsasymptotic stabilityobservabilityuniform approximationnonlinear filteringmerging of probability measures

Mathematics Subject Classification ID

Inference from stochastic processes and prediction (62M20) Filtering in stochastic control theory (93E11) Signal theory (characterization, reconstruction, filtering, etc.) (94A12) Continuous-time Markov processes on general state spaces (60J25) Observability (93B07) Stochastic stability in control theory (93E15)

Cites Work

• Title not available (Why is that?)
• Real Analysis and Probability
• On the consistency of Bayes estimates
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Merging of Opinions with Increasing Information
• Title not available (Why is that?)
• Stability and Uniform Particle Approximation of Nonlinear Filters in Case of Non Ergodic Signals
• Title not available (Why is that?)
• Positive operators
• Markov additive processes. I
• Martingale problems for conditional distributions of Markov processes
• A robustification approach to stability and to uniform particle approximation of nonlinear filters: the example of pseudo-mixing signals.
• Convergence of Baire measures
• The stability of conditional Markov processes and Markov chains in random environments
• An Operator Theorem on $L_1$ Convergence to Zero with Applications to Markov Kernels
• Asymptotic Stability of the Optimal Filter with Respect to Its Initial Condition
• Observability and nonlinear filtering
• Exponential stability of discrete-time filters for bounded observation noise
• Exponential stability in discrete-time filtering for non-ergodic signal.
• Stability of the discrete time filter in terms of the tails of noise distributions
• Measures as functionals on uniformly continuous functions
• Title not available (Why is that?)
• Stability of the filter equation for a time-dependent signal on \(\mathbb R^{d}\)
• On a role of predictor in the filtering stability
• Title not available (Why is that?)
• Uniform Approximation on Noncompact Spaces

Cited In (13)

• Uniform observability and exponential convergence rate of the Kalman filter for the FIR deconvolution problem.
• Exponential stability of filters and smoothers for hidden Markov models
• Uniform time average consistency of Monte Carlo particle filters
• Sequential Monte Carlo smoothing for general state space hidden Markov models
• Asymptotic behavior of the forecast-assimilation process with unstable dynamics
• Stability properties of systems of linear stochastic differential equations with random coefficients
• A note on uniform observability
• A complete solution to Blackwell's unique ergodicity problem for hidden Markov chains
• Exponential forgetting of smoothing distributions for pairwise Markov models
• The mean-field ensemble Kalman filter: near-Gaussian setting
• Robustness to incorrect priors and controlled filter stability in partially observed stochastic control
• Planning and learning in partially observable systems via filter stability
• Forgetting of the initial distribution for nonergodic hidden Markov chains

This page was built for publication: Uniform observability of hidden Markov models and filter stability for unstable signals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2389607)