Lasso-Kalman smoother for tracking sparse signals

Research output: Chapter in Book/Report/Conference proceedingConference contribution

29 Scopus citations

Abstract

Fixed-interval smoothing of time-varying vector processes is an estimation approach with well-documented merits for tracking applications. The optimal performance in the linear Gauss-Markov model is achieved by the Kalman smoother (KS), which also admits an efficient recursive implementation. The present paper deals with vector processes for which it is known a priori that many of their entries equal to zero. In this context, the process to be tracked is sparse, and the performance of sparsity-agnostic KS schemes degrades considerably. On the other hand, it is shown here that a sparsity-aware KS exhibits complexity which grows exponentially in the vector dimension. To obtain a tractable alternative, the KS cost is regularized with the sparsity-promoting l1 norm of the vector process - a relaxation also used in linear regression problems to obtain the least-absolute shrinkage and selection operator (Lasso). The Lasso (L)KS derived in this work is not only capable of tracking sparse time-varying vector processes, but can also afford an efficient recursive implementation based on the alternating direction method of multipliers (ADMoM). Finally, a weighted (W)-LKS is also introduced to cope with the bias of the LKS, and simulations are provided to validate the performance of the novel algorithms

Original languageEnglish (US)
Title of host publicationConference Record - 43rd Asilomar Conference on Signals, Systems and Computers
Pages181-185
Number of pages5
DOIs
StatePublished - 2009
Event43rd Asilomar Conference on Signals, Systems and Computers - Pacific Grove, CA, United States
Duration: Nov 1 2009Nov 4 2009

Publication series

NameConference Record - Asilomar Conference on Signals, Systems and Computers
ISSN (Print)1058-6393

Other

Other43rd Asilomar Conference on Signals, Systems and Computers
Country/TerritoryUnited States
CityPacific Grove, CA
Period11/1/0911/4/09

Fingerprint

Dive into the research topics of 'Lasso-Kalman smoother for tracking sparse signals'. Together they form a unique fingerprint.

Cite this