© University of Kent - Contact | Feedback | Legal | FOI | Cookies
Kullback-Leibler approach to Gaussian mixture reduction
Andrew R. Runnalls
IEEE Transactions of Aerospace and Electronic Systems, 43(3):182-196, July 2007.Abstract
A Kullback-Leibler Approach to Gaussian Mixture Reduction
Andrew R. Runnalls University of Kent
A common problem in multi-target tracking is to approximate a Gaussian mixture by one containing fewer components; similar problems can arise in integrated navigation. A common approach is successively to merge pairs of components, replacing the pair with a single Gaussian component whose moments up to second order match those of the merged pair. Salmond [1990] and Williams [2003, 2003a] have each proposed algorithms along these lines, but using different criteria for selecting the pair to be merged at each stage. The paper shows how under certain circumstances each of these pair-selection criteria can give rise to anomalous behaviour, and proposes that a key consideration should be the Kullback-Leibler discrimination of the reduced mixture with respect to the original mixture. Although computing this directly would normally be impractical, the paper shows how an easily-computed upper bound can be used as a pair-selection criterion which avoids the anomalies of the earlier approaches. The behaviour of the three algorithms is compared using a high-dimensional example drawn from terrain-referenced navigation.
Download publication 254 kbytes (PDF)Bibtex Record
@article{2797,
author = {Andrew R. Runnalls},
title = {Kullback-{L}eibler Approach to {G}aussian Mixture Reduction},
month = {July},
year = {2007},
pages = {182-196},
keywords = {determinacy analysis, Craig interpolants},
note = {},
doi = {},
url = {http://www.cs.kent.ac.uk/pubs/2007/2797},
publication_type = {article},
submission_id = {23453_1219928502},
ISSN = {0018-9251},
journal = {IEEE Transactions of Aerospace and Electronic Systems},
volume = {43},
number = {3},
publisher = {IEEE},
}