Pathwise nonlinear filtering on abstract Wiener spaces

DOI10.1214/AOP/1176989139MaRDI QIDQ688077zbMATH OpenOpenAlexFDO

Authors Andrew Lyasoff

Publication date 21 July 1994

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

Full work available at URL https://doi.org/10.1214/aop/1176989139

zbMATH Keywords

finite energy conditionGirsanov formulaKallianpur-Striebel formulanonlinear filtering theory on abstract Wiener spacesZakai equation

Mathematics Subject Classification ID

Gaussian processes (60G15) Signal detection and filtering (aspects of stochastic processes) (60G35) Stochastic integral equations (60H20) Filtering in stochastic control theory (93E11) Signal theory (characterization, reconstruction, filtering, etc.) (94A12)

Review text

This paper is concerned with the nonlinear filtering theory on abstract Wiener spaces. The system takes the form \(y=h+\omega\) where \(h\) runs in a separable Hilbert space \(H\) and \(\omega\) runs through some Banach space \(E\). \(H\) is densely embedded in \(E\) and therefore, they form an abstract Wiener space. Considering shift transformations on \(E\), the author establishes an analog of the Girsanov formula. Then, a Kallianpur- Striebel formula for signal correlated with noise is obtained. This allows to express the filter with an explicit functional of the observed trajectory. Under a finite energy condition, the author proves also a pathwise version of the Zakai equation.

Recommendations

• scientific article; zbMATH DE number 4095357
• scientific article; zbMATH DE number 1370377
• scientific article; zbMATH DE number 3906370
• scientific article; zbMATH DE number 956695
• On filtering equations in infinite dimensions

Cited in

(3)

• A mapping result between Wiener theory and Kalman filtering for nonstationary processes
• Path continuity of the nonlinear filter
• The construction of filtrations on abstract Wiener space

This page was built for publication: Pathwise nonlinear filtering on abstract Wiener spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q688077)