Sparsity-aware Kalman tracking of target signal strengths on a grid

Shahrokh Farahmand, Georgios B. Giannakis, Geert Leus, Zhi Tian

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

10 Scopus citations

Abstract

Tracking multiple moving targets is known to be challenged by the nonlinearity present in the measurement equation, and by the computationally burdensome data association task. This paper introduces a grid-based model of target signal strengths leading to linear state and measurement equations, that can afford state estimation via sparsity-aware Kalman filtering (KF), and bypasses data association. Leveraging the sparsity inherent to the novel grid-based model, a sparsity-cognizant KF tracker is developed that effects sparsity through ℓ1-norm regularization. The proposed tracker does not require knowledge of the number of targets or their signal strengths, and exhibits considerably lower complexity than the hidden Markov filter benchmark, especially as the number of targets increases. Numerical simulations demonstrate that the sparsity-cognizant tracker enjoys improved root mean-square error performance at reduced complexity when compared to its sparsity-agnostic counterparts.

Original languageEnglish (US)
Title of host publicationFusion 2011 - 14th International Conference on Information Fusion
StatePublished - Sep 13 2011
Event14th International Conference on Information Fusion, Fusion 2011 - Chicago, IL, United States
Duration: Jul 5 2011Jul 8 2011

Publication series

NameFusion 2011 - 14th International Conference on Information Fusion

Conference

Conference14th International Conference on Information Fusion, Fusion 2011
CountryUnited States
CityChicago, IL
Period7/5/117/8/11

Keywords

  • Compressed sensing
  • Kalman filter
  • Multi-target tracking
  • Sparsity

Fingerprint Dive into the research topics of 'Sparsity-aware Kalman tracking of target signal strengths on a grid'. Together they form a unique fingerprint.

Cite this