A CPHD approximation based on a discrete-Gamma cardinality model

Flavio Eler De Melo, Simon Maskell

Research output: Contribution to journalArticle

Abstract

The Cardinalized Probability Hypothesis Density (CPHD) filter has become one of the most acclaimed algorithms for multi-target Bayesian filtering due to its ability to accurately estimate the number of objects and the object states in tracking scenarios affected by clutter. The CPHD filter generalizes the Probabilistic Hypothesis Density (PHD) filter by jointly propa- gating the first-order multi-target moment (intensity function) along with the entire probability distribution on the number of targets (cardinality distribution). In general, the CPHD recursion is computationally intractable, however successful approxima- tions have been devised with reported computational complexity dominated by O(m3) operations per filtering iteration, where m is the number of measurements. Room for improvement was originally acknowledged by Mahler, who conceived the idea of approximating the cardinality distribution by two-parameter distributions. In this paper, we further explore this idea to provide an efficient approximation of the CPHD filter where the cardinality distribution is modeled as a discretized Gamma distribution. Experiments show that the resulting filter is less computationally complex than the standard implementation of the CPHD filter but shows similar cardinality accuracy and variance.
LanguageEnglish
JournalIEEE Transactions on Signal Processing
Early online date16 Nov 2018
DOIs
StateE-pub ahead of print - 16 Nov 2018

Fingerprint

Probability distributions
Computational complexity
Experiments

Keywords

  • Approximation algorithms
  • Bayes methods
  • Probabilistic logic
  • Radio frequency
  • Signal processing algorithms
  • Target tracking

Cite this

@article{d345f98ba67f40449bbaa60def136f3f,
title = "A CPHD approximation based on a discrete-Gamma cardinality model",
abstract = "The Cardinalized Probability Hypothesis Density (CPHD) filter has become one of the most acclaimed algorithms for multi-target Bayesian filtering due to its ability to accurately estimate the number of objects and the object states in tracking scenarios affected by clutter. The CPHD filter generalizes the Probabilistic Hypothesis Density (PHD) filter by jointly propa- gating the first-order multi-target moment (intensity function) along with the entire probability distribution on the number of targets (cardinality distribution). In general, the CPHD recursion is computationally intractable, however successful approxima- tions have been devised with reported computational complexity dominated by O(m3) operations per filtering iteration, where m is the number of measurements. Room for improvement was originally acknowledged by Mahler, who conceived the idea of approximating the cardinality distribution by two-parameter distributions. In this paper, we further explore this idea to provide an efficient approximation of the CPHD filter where the cardinality distribution is modeled as a discretized Gamma distribution. Experiments show that the resulting filter is less computationally complex than the standard implementation of the CPHD filter but shows similar cardinality accuracy and variance.",
keywords = "Approximation algorithms, Bayes methods, Probabilistic logic, Radio frequency, Signal processing algorithms, Target tracking",
author = "{De Melo}, {Flavio Eler} and Simon Maskell",
year = "2018",
month = "11",
day = "16",
doi = "10.1109/TSP.2018.2881659",
language = "English",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "IEEE",

}

A CPHD approximation based on a discrete-Gamma cardinality model. / De Melo, Flavio Eler; Maskell, Simon.

In: IEEE Transactions on Signal Processing, 16.11.2018.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A CPHD approximation based on a discrete-Gamma cardinality model

AU - De Melo,Flavio Eler

AU - Maskell,Simon

PY - 2018/11/16

Y1 - 2018/11/16

N2 - The Cardinalized Probability Hypothesis Density (CPHD) filter has become one of the most acclaimed algorithms for multi-target Bayesian filtering due to its ability to accurately estimate the number of objects and the object states in tracking scenarios affected by clutter. The CPHD filter generalizes the Probabilistic Hypothesis Density (PHD) filter by jointly propa- gating the first-order multi-target moment (intensity function) along with the entire probability distribution on the number of targets (cardinality distribution). In general, the CPHD recursion is computationally intractable, however successful approxima- tions have been devised with reported computational complexity dominated by O(m3) operations per filtering iteration, where m is the number of measurements. Room for improvement was originally acknowledged by Mahler, who conceived the idea of approximating the cardinality distribution by two-parameter distributions. In this paper, we further explore this idea to provide an efficient approximation of the CPHD filter where the cardinality distribution is modeled as a discretized Gamma distribution. Experiments show that the resulting filter is less computationally complex than the standard implementation of the CPHD filter but shows similar cardinality accuracy and variance.

AB - The Cardinalized Probability Hypothesis Density (CPHD) filter has become one of the most acclaimed algorithms for multi-target Bayesian filtering due to its ability to accurately estimate the number of objects and the object states in tracking scenarios affected by clutter. The CPHD filter generalizes the Probabilistic Hypothesis Density (PHD) filter by jointly propa- gating the first-order multi-target moment (intensity function) along with the entire probability distribution on the number of targets (cardinality distribution). In general, the CPHD recursion is computationally intractable, however successful approxima- tions have been devised with reported computational complexity dominated by O(m3) operations per filtering iteration, where m is the number of measurements. Room for improvement was originally acknowledged by Mahler, who conceived the idea of approximating the cardinality distribution by two-parameter distributions. In this paper, we further explore this idea to provide an efficient approximation of the CPHD filter where the cardinality distribution is modeled as a discretized Gamma distribution. Experiments show that the resulting filter is less computationally complex than the standard implementation of the CPHD filter but shows similar cardinality accuracy and variance.

KW - Approximation algorithms

KW - Bayes methods

KW - Probabilistic logic

KW - Radio frequency

KW - Signal processing algorithms

KW - Target tracking

UR - http://www.scopus.com/inward/record.url?scp=85056739303&partnerID=8YFLogxK

U2 - 10.1109/TSP.2018.2881659

DO - 10.1109/TSP.2018.2881659

M3 - Article

JO - IEEE Transactions on Signal Processing

T2 - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

ER -