1976•Unpublished venueRequires access

Levinson- and chandrasekhar-type equations for a general discrete-time linear estimation problem

B. Friedlander, Martin E. Morf, T. Kailath, Lennart Ljung

Open publisher page 5 citations

Abstract

Recursive algorithms for the solution of linear least-squares estimation problems have been based mainly on state-space models. It has been know, however, that such algorithms exist for stationary time-series, using input-output descriptions (e.g., covariance matrices). We introduce a way of classifying stochastic processes in terms of their "distance" from stationarity that leads to a derivation of an efficient Levinson-type algorithm for arbitrary (nonstationary) processes. By adding structure to the covariance matrix, these general results specialize to state-space type estimation algorithms. In particular, the Chandrasekhar equations are shown to be the natural descendants of the Levinson algorithm.

About this research paper

What this paper is about

Recursive algorithms for the solution of linear least-squares estimation problems have been based mainly on state-space models. It has been know, however, that such algorithms exist for stationary time-series, using input-output descriptions (e.g., covariance matrices). We introduce a way of classifying stochastic processes in terms of their "distance" from stationarity that leads to a derivation of an efficient Levinson-type algorithm for arbitrary (nonstationary) processes. By adding structure to the covariance matrix, these general results specialize to state-space type estimation algorithms. In particular, the Chandrasekhar equations are shown to be the natural descendants of the Levinson algorithm.

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Recursive algorithms for the solution of linear least-squares estimation problems have been based mainly on state-space models. It has been know, however, that such algorithms exist for stationary time-series, using input-output descriptions (e.g., covariance matrices). We introduce a way of classifying stochastic processes in terms of their "distance" from stationarity that leads to a derivation of an efficient Levinson-type algorithm for arbitrary (nonstationary) processes. By adding structure to the covariance matrix, these general results specialize to state-space type estimation algorithms. In particular, the Chandrasekhar equations are shown to be the natural descendants of the Levinson algorithm.

Key concepts: Chandrasekhar limit, Covariance, Type (biology), Applied mathematics, Mathematics, Covariance matrix, State space, Series (stratigraphy)

Related papers

Back to paper searchBrowse research topicsOriginal source
Levinson- and chandrasekhar-type equations for a general discrete-time linear estimation problem — Research Paper | ScholarLens