A Variational Measurement Update for Extended Target Tracking With Random Matrices

Orguner U.

IEEE TRANSACTIONS ON SIGNAL PROCESSING, vol.60, no.7, pp.3827-3834, 2012 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 60 Issue: 7
  • Publication Date: 2012
  • Doi Number: 10.1109/tsp.2012.2192927
  • Journal Name: IEEE TRANSACTIONS ON SIGNAL PROCESSING
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.3827-3834
  • Keywords: Extended target tracking, measurement update, random matrices, variational Bayes, OBJECT
  • Middle East Technical University Affiliated: Yes

Abstract

This correspondence proposes a new measurement update for extended target tracking under measurement noise when the target extent is modeled by random matrices. Compared to the previous measurement update developed by Feldmann et al., this work follows a more rigorous path to derive an approximate measurement update using the analytical techniques of variational Bayesian inference. The resulting measurement update, though computationally more expensive, is shown via simulations to be better than the earlier method in terms of both the state estimates and the predictive likelihood for moderate amounts of prediction errors.