Multi-target FIR tracking algorithm for Markov jump linear systems based on true-target decision-making

Chang Joo Lee; Jung Min Pak; Choon Ki Ahn; Kyung Min Min; Peng Shi; Myo Taeg Lim

doi:10.1016/j.neucom.2015.05.096

Multi-target FIR tracking algorithm for Markov jump linear systems based on true-target decision-making

Chang Joo Lee, Jung Min Pak, Choon Ki Ahn, Kyung Min Min, Peng Shi, Myo Taeg Lim

Research output: Contribution to journal › Article › peer-review

22 Citations (Scopus)

Abstract

Most existing multi-target tracking (MTT) algorithms are based on Kalman filters (KFs). However, KFs exhibit poor estimation performance or even diverge when system models have parameter uncertainties. To overcome this drawback, finite impulse response (FIR) filters have been studied; these are more robust against model uncertainty than KFs. In this paper, we propose a novel MTT algorithm based on FIR filtering for Markov jump linear systems (MJLSs). The proposed algorithm is called the multi-target FIR tracking algorithm (MTFTA). The MTFTA is based on the decision-making process to identify the true-target[U+05F3]s state among candidate states. The true-target decision-making process utilizes the likelihood function and the Mahalanobis distance. We show that the proposed MTFTA exhibits better robustness against model parameter uncertainties than the conventional KF-based algorithm.

Original language	English
Pages (from-to)	298-307
Number of pages	10
Journal	Neurocomputing
Volume	168
DOIs	https://doi.org/10.1016/j.neucom.2015.05.096
Publication status	Published - 2015 Nov 30

Bibliographical note

Funding Information:
This work was supported partially by the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning ( NRF-2014R1A1A1006101 ), partially by General Research Program through National Research Foundation of Korea funded by the Ministry of Education (Grant no. NRF-2013R1A1A2008698 ), partially by the Australian Research Council ( DP140102180 , LP140100471 , LE150100079 ), and the 111 Project (B12018) .

Publisher Copyright:
© 2015 Elsevier B.V.

Keywords

Finite impulse response (FIR) filter
Markov jump linear system (MJLS)
Multi-target FIR tracking algorithm (MTFTA)
Multi-target tracking (MTT)
True-target decision making

ASJC Scopus subject areas

Computer Science Applications
Cognitive Neuroscience
Artificial Intelligence

Access to Document

10.1016/j.neucom.2015.05.096

Cite this

@article{7c6b3239a7e342bc8879427cc2c2c708,

title = "Multi-target FIR tracking algorithm for Markov jump linear systems based on true-target decision-making",

abstract = "Most existing multi-target tracking (MTT) algorithms are based on Kalman filters (KFs). However, KFs exhibit poor estimation performance or even diverge when system models have parameter uncertainties. To overcome this drawback, finite impulse response (FIR) filters have been studied; these are more robust against model uncertainty than KFs. In this paper, we propose a novel MTT algorithm based on FIR filtering for Markov jump linear systems (MJLSs). The proposed algorithm is called the multi-target FIR tracking algorithm (MTFTA). The MTFTA is based on the decision-making process to identify the true-target[U+05F3]s state among candidate states. The true-target decision-making process utilizes the likelihood function and the Mahalanobis distance. We show that the proposed MTFTA exhibits better robustness against model parameter uncertainties than the conventional KF-based algorithm.",

keywords = "Finite impulse response (FIR) filter, Markov jump linear system (MJLS), Multi-target FIR tracking algorithm (MTFTA), Multi-target tracking (MTT), True-target decision making",

author = "Lee, {Chang Joo} and Pak, {Jung Min} and Ahn, {Choon Ki} and Min, {Kyung Min} and Peng Shi and Lim, {Myo Taeg}",

note = "Funding Information: This work was supported partially by the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning ( NRF-2014R1A1A1006101 ), partially by General Research Program through National Research Foundation of Korea funded by the Ministry of Education (Grant no. NRF-2013R1A1A2008698 ), partially by the Australian Research Council ( DP140102180 , LP140100471 , LE150100079 ), and the 111 Project (B12018) . Publisher Copyright: {\textcopyright} 2015 Elsevier B.V.",

year = "2015",

month = nov,

day = "30",

doi = "10.1016/j.neucom.2015.05.096",

language = "English",

volume = "168",

pages = "298--307",

journal = "Neurocomputing",

issn = "0925-2312",

publisher = "Elsevier",

}

TY - JOUR

T1 - Multi-target FIR tracking algorithm for Markov jump linear systems based on true-target decision-making

AU - Lee, Chang Joo

AU - Pak, Jung Min

AU - Ahn, Choon Ki

AU - Min, Kyung Min

AU - Shi, Peng

AU - Lim, Myo Taeg

N1 - Funding Information: This work was supported partially by the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning ( NRF-2014R1A1A1006101 ), partially by General Research Program through National Research Foundation of Korea funded by the Ministry of Education (Grant no. NRF-2013R1A1A2008698 ), partially by the Australian Research Council ( DP140102180 , LP140100471 , LE150100079 ), and the 111 Project (B12018) . Publisher Copyright: © 2015 Elsevier B.V.

PY - 2015/11/30

Y1 - 2015/11/30

N2 - Most existing multi-target tracking (MTT) algorithms are based on Kalman filters (KFs). However, KFs exhibit poor estimation performance or even diverge when system models have parameter uncertainties. To overcome this drawback, finite impulse response (FIR) filters have been studied; these are more robust against model uncertainty than KFs. In this paper, we propose a novel MTT algorithm based on FIR filtering for Markov jump linear systems (MJLSs). The proposed algorithm is called the multi-target FIR tracking algorithm (MTFTA). The MTFTA is based on the decision-making process to identify the true-target[U+05F3]s state among candidate states. The true-target decision-making process utilizes the likelihood function and the Mahalanobis distance. We show that the proposed MTFTA exhibits better robustness against model parameter uncertainties than the conventional KF-based algorithm.

AB - Most existing multi-target tracking (MTT) algorithms are based on Kalman filters (KFs). However, KFs exhibit poor estimation performance or even diverge when system models have parameter uncertainties. To overcome this drawback, finite impulse response (FIR) filters have been studied; these are more robust against model uncertainty than KFs. In this paper, we propose a novel MTT algorithm based on FIR filtering for Markov jump linear systems (MJLSs). The proposed algorithm is called the multi-target FIR tracking algorithm (MTFTA). The MTFTA is based on the decision-making process to identify the true-target[U+05F3]s state among candidate states. The true-target decision-making process utilizes the likelihood function and the Mahalanobis distance. We show that the proposed MTFTA exhibits better robustness against model parameter uncertainties than the conventional KF-based algorithm.

KW - Finite impulse response (FIR) filter

KW - Markov jump linear system (MJLS)

KW - Multi-target FIR tracking algorithm (MTFTA)

KW - Multi-target tracking (MTT)

KW - True-target decision making

UR - http://www.scopus.com/inward/record.url?scp=84937811700&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2015.05.096

DO - 10.1016/j.neucom.2015.05.096

M3 - Article

AN - SCOPUS:84937811700

SN - 0925-2312

VL - 168

SP - 298

EP - 307

JO - Neurocomputing

JF - Neurocomputing

ER -

Multi-target FIR tracking algorithm for Markov jump linear systems based on true-target decision-making

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this