Premium
Stability of the discrete time filter in terms of the tails of noise distributions
Author(s) -
Crisan D.,
Heine K.
Publication year - 2008
Publication title -
journal of the london mathematical society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.441
H-Index - 62
eISSN - 1469-7750
pISSN - 0024-6107
DOI - 10.1112/jlms/jdn032
Subject(s) - stability (learning theory) , mathematics , filter (signal processing) , ergodicity , state space , noise (video) , control theory (sociology) , randomness , convergence (economics) , discrete time and continuous time , statistical physics , computer science , physics , statistics , control (management) , machine learning , artificial intelligence , economics , image (mathematics) , computer vision , economic growth
In recent years, the stability of discrete time filters has been a field of active research. By stability we mean that the effect of the possibly erroneous initial distribution in the filter eventually vanishes as time increases. One of the motivations for our interest in the stability is its close relation to the convergence of various numerical filter approximation schemes, for example, particle filters. In this paper, the main result states easily verifiable conditions that are sufficient for filter stability. Essentially, the conditions state that the filter is stable, if the observation noise is sufficiently light tailed compared with the randomness in the signal process. Compactness of the state space or ergodicity of the signal is not required.