Please use this identifier to cite or link to this item:
https://hdl.handle.net/2440/82887
Citations | ||
Scopus | Web of Science® | Altmetric |
---|---|---|
?
|
?
|
Type: | Conference paper |
Title: | H-PMHT with a poisson measurement model |
Author: | Vu, H. Davey, S. Arulampalam, M. Fletcher, F. Lim, C. |
Citation: | Proceedings of the International Conference on Radar, RADAR 2013, 9-12 September 2013, Adelaide, Australia: pp.446-451 |
Publisher: | IEEE |
Publisher Place: | USA/CD |
Issue Date: | 2013 |
ISBN: | 9781467351782 |
Conference Name: | International Conference on Radar (2013 : Adelaide, Australia) |
Statement of Responsibility: | Han X. Vu, Samuel J. Davey, Sanjeev Arulampalam, Fiona K. Fletcher and Cheng-Chew Lim |
Abstract: | The Histogram Probabilistic Multi-Hypothesis Tracker (H-PMHT) is an efficient multi-target approach to the Track-before-detect (TkBD) problem. The tracking is based on the generation of a synthetic histogram by quantising the energy in the sensor data. The resultant quantised measurement is then modelled using a multinomial distribution and target state estimation is performed via Expectation-Maximisation based mixture modeling. This paper presents an alternative derivation of the H-PMHT based on a Poisson measurement model. The benefits of this new derivation are two-fold. First, direct estimation of the measurement likelihood is now possible under this new formulation, thereby eliminating any need for measurement quantisation. Second, the new derivation results in an improved measure for track quality by incorporating a time-correlated estimate for the target mixing proportions. |
Rights: | © 2013 Commonwealth of Australia |
DOI: | 10.1109/RADAR.2013.6652030 |
Published version: | http://dx.doi.org/10.1109/radar.2013.6652030 |
Appears in Collections: | Aurora harvest 4 Electrical and Electronic Engineering publications |
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.