Probabilistic Labeling in Radar Track-before-Detect Processing: Algorithms for tracking closely-spaced and/or interacting targets

C. Moreno León

doi:10.4233/uuid:039445ea-ce0e-445d-b07c-5e369fe7e708

Probabilistic Labeling in Radar Track-before-Detect Processing: Algorithms for tracking closely-spaced and/or interacting targets

C. Moreno León

Microwave Sensing, Signals & Systems

Research output: Thesis › Dissertation (TU Delft)

29 Downloads (Pure)

Abstract

Radar-tracking of low-observable targets such as drones suffers from low detection performance. In these type of applications, it is desirable to avoid data thresholding in order to preserve the weak target signal in the raw sensor data. This thesis considers the Multiple Object Tracking (MOT) problem in the context of radar Track-before-Detect (TrBD) processing, where the raw radar data is fed into the filtering process without previous compression into a finite set of detection/plots....

Original language	English
Qualification	Doctor of Philosophy
Awarding Institution	Delft University of Technology
Supervisors/Advisors	Yarovoy , Alexander, Supervisor Driessen, J.N., Advisor
Award date	27 Nov 2023
Print ISBNs	978-94-6384-512-0
DOIs	https://doi.org/10.4233/uuid:039445ea-ce0e-445d-b07c-5e369fe7e708
Publication status	Published - 2023

Keywords

Multiple target tracking
Radar Track-before-detect
Bayesian Inference
Tracking of interacting/closely-spaced/unresolved targets
non-linear filtering
detection of target anomalous behaviour
Particle filtering
data-association free tracking
Sequential Monte Carlo methods

Access to Document

10.4233/uuid:039445ea-ce0e-445d-b07c-5e369fe7e708

dissertation_Carlos_MorenoFinal published version, 10.1 MB

Cite this

@phdthesis{039445eace0e445db07c5e369fe7e708,

title = "Probabilistic Labeling in Radar Track-before-Detect Processing: Algorithms for tracking closely-spaced and/or interacting targets",

abstract = "Radar-tracking of low-observable targets such as drones suffers from low detection performance. In these type of applications, it is desirable to avoid data thresholding in order to preserve the weak target signal in the raw sensor data. This thesis considers the Multiple Object Tracking (MOT) problem in the context of radar Track-before-Detect (TrBD) processing, where the raw radar data is fed into the filtering process without previous compression into a finite set of detection/plots....",

keywords = "Multiple target tracking, Radar Track-before-detect, Bayesian Inference, Tracking of interacting/closely-spaced/unresolved targets, non-linear filtering, detection of target anomalous behaviour, Particle filtering, data-association free tracking, Sequential Monte Carlo methods",

author = "{Moreno Le{\'o}n}, C.",

year = "2023",

doi = "10.4233/uuid:039445ea-ce0e-445d-b07c-5e369fe7e708",

language = "English",

isbn = "978-94-6384-512-0",

type = "Dissertation (TU Delft)",

school = "Delft University of Technology",

}

TY - THES

T1 - Probabilistic Labeling in Radar Track-before-Detect Processing

T2 - Algorithms for tracking closely-spaced and/or interacting targets

AU - Moreno León, C.

PY - 2023

Y1 - 2023

N2 - Radar-tracking of low-observable targets such as drones suffers from low detection performance. In these type of applications, it is desirable to avoid data thresholding in order to preserve the weak target signal in the raw sensor data. This thesis considers the Multiple Object Tracking (MOT) problem in the context of radar Track-before-Detect (TrBD) processing, where the raw radar data is fed into the filtering process without previous compression into a finite set of detection/plots....

AB - Radar-tracking of low-observable targets such as drones suffers from low detection performance. In these type of applications, it is desirable to avoid data thresholding in order to preserve the weak target signal in the raw sensor data. This thesis considers the Multiple Object Tracking (MOT) problem in the context of radar Track-before-Detect (TrBD) processing, where the raw radar data is fed into the filtering process without previous compression into a finite set of detection/plots....

KW - Multiple target tracking

KW - Radar Track-before-detect

KW - Bayesian Inference

KW - Tracking of interacting/closely-spaced/unresolved targets

KW - non-linear filtering

KW - detection of target anomalous behaviour

KW - Particle filtering

KW - data-association free tracking

KW - Sequential Monte Carlo methods

U2 - 10.4233/uuid:039445ea-ce0e-445d-b07c-5e369fe7e708

DO - 10.4233/uuid:039445ea-ce0e-445d-b07c-5e369fe7e708

M3 - Dissertation (TU Delft)

SN - 978-94-6384-512-0

ER -

Probabilistic Labeling in Radar Track-before-Detect Processing: Algorithms for tracking closely-spaced and/or interacting targets

Abstract

Keywords

Access to Document

Fingerprint

Cite this