Accelerating the single cluster PHD filter with a GPU implementation

Chee Sing Lee; Jose Franco; Jeremie Houssineau; Daniel Clark

doi:10.1109/ICCAIS.2014.7020567

Accelerating the single cluster PHD filter with a GPU implementation

Chee Sing Lee^*, Jose Franco, Jeremie Houssineau, Daniel Clark

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Citations (Scopus)

Abstract

The SC-PHD filter is an algorithm which was designed to solve a class of multiple object estimation problems where it is necessary to estimate the state of a single-target parent process, in addition to estimating the state of a multi-object population which is conditioned on it. The filtering process usually employs a number of particles to represent the parent process, coupled each with a conditional PHD filter, which is computationally burdensome. In this article, an implementation is described which exploits the parallel nature of the filter to obtain considerable speed-up with the help of a GPU. Several considerations need to be taken into account to make efficient use of the GPU, and these are also described here.

Original language	English
Title of host publication	2014 International Conference on Control, Automation and Information Sciences, ICCAIS 2014
Publisher	IEEE
Pages	53-58
Number of pages	6
ISBN (Print)	9781479972043
DOIs	https://doi.org/10.1109/ICCAIS.2014.7020567
Publication status	Published - 2014
Event	3rd International Conference on Control, Automation and Information Sciences - Gwangju, United Kingdom Duration: 2 Dec 2014 → 5 Dec 2014

Conference

Conference	3rd International Conference on Control, Automation and Information Sciences
Abbreviated title	ICCAIS 2014
Country/Territory	United Kingdom
City	Gwangju
Period	2/12/14 → 5/12/14

Access to Document

10.1109/ICCAIS.2014.7020567

Cite this

@inproceedings{5686fd071fff490699fedcb514711453,

title = "Accelerating the single cluster PHD filter with a GPU implementation",

abstract = "The SC-PHD filter is an algorithm which was designed to solve a class of multiple object estimation problems where it is necessary to estimate the state of a single-target parent process, in addition to estimating the state of a multi-object population which is conditioned on it. The filtering process usually employs a number of particles to represent the parent process, coupled each with a conditional PHD filter, which is computationally burdensome. In this article, an implementation is described which exploits the parallel nature of the filter to obtain considerable speed-up with the help of a GPU. Several considerations need to be taken into account to make efficient use of the GPU, and these are also described here.",

author = "Lee, \{Chee Sing\} and Jose Franco and Jeremie Houssineau and Daniel Clark",

year = "2014",

doi = "10.1109/ICCAIS.2014.7020567",

language = "English",

isbn = "9781479972043",

pages = "53--58",

booktitle = "2014 International Conference on Control, Automation and Information Sciences, ICCAIS 2014",

publisher = "IEEE",

address = "United States",

note = "3rd International Conference on Control, Automation and Information Sciences, ICCAIS 2014 ; Conference date: 02-12-2014 Through 05-12-2014",

}

Lee, CS, Franco, J, Houssineau, J & Clark, D 2014, Accelerating the single cluster PHD filter with a GPU implementation. in 2014 International Conference on Control, Automation and Information Sciences, ICCAIS 2014., 7020567, IEEE, pp. 53-58, 3rd International Conference on Control, Automation and Information Sciences, Gwangju, United Kingdom, 2/12/14. https://doi.org/10.1109/ICCAIS.2014.7020567

TY - GEN

T1 - Accelerating the single cluster PHD filter with a GPU implementation

AU - Lee, Chee Sing

AU - Franco, Jose

AU - Houssineau, Jeremie

AU - Clark, Daniel

PY - 2014

Y1 - 2014

N2 - The SC-PHD filter is an algorithm which was designed to solve a class of multiple object estimation problems where it is necessary to estimate the state of a single-target parent process, in addition to estimating the state of a multi-object population which is conditioned on it. The filtering process usually employs a number of particles to represent the parent process, coupled each with a conditional PHD filter, which is computationally burdensome. In this article, an implementation is described which exploits the parallel nature of the filter to obtain considerable speed-up with the help of a GPU. Several considerations need to be taken into account to make efficient use of the GPU, and these are also described here.

AB - The SC-PHD filter is an algorithm which was designed to solve a class of multiple object estimation problems where it is necessary to estimate the state of a single-target parent process, in addition to estimating the state of a multi-object population which is conditioned on it. The filtering process usually employs a number of particles to represent the parent process, coupled each with a conditional PHD filter, which is computationally burdensome. In this article, an implementation is described which exploits the parallel nature of the filter to obtain considerable speed-up with the help of a GPU. Several considerations need to be taken into account to make efficient use of the GPU, and these are also described here.

U2 - 10.1109/ICCAIS.2014.7020567

DO - 10.1109/ICCAIS.2014.7020567

M3 - Conference contribution

SN - 9781479972043

SP - 53

EP - 58

BT - 2014 International Conference on Control, Automation and Information Sciences, ICCAIS 2014

PB - IEEE

T2 - 3rd International Conference on Control, Automation and Information Sciences

Y2 - 2 December 2014 through 5 December 2014

ER -

Accelerating the single cluster PHD filter with a GPU implementation

Abstract

Conference

Access to Document

Fingerprint

Cite this