A Graph Neural Network Approach for Solving the Ranked Assignment Problem in Multi-Object Tracking

Loading...
Thumbnail Image

Date

2024-07-15

Authors

Dehler Robin
Herrmann Martin
Strohbeck Jan
Buchholz Michael

Journal Title

Journal ISSN

Volume Title

Publication Type

Beitrag zu einer Konferenz

Published in

2024 IEEE Intelligent Vehicles Symposium (IV), 2024

Abstract

Associating measurements with tracks is a crucial step in Multi-Object Tracking (MOT) to guarantee the safety of autonomous vehicles. To manage the exponentially growing number of track hypotheses, truncation becomes necessary. In the δ-Generalized Labeled Multi-Bernoulli (δ-GLMB) filter application, this truncation typically involves the ranked assignment problem, solved by Murty’s algorithm or the Gibbs sampling approach, both with limitations in terms of complexity or accuracy, respectively. With the motivation to improve these limitations, this paper addresses the ranked assignment problem arising from data association tasks with an approach that employs Graph Neural Networks (GNNs). The proposed Ranked Assignment Prediction Graph Neural Network (RAPNet) uses bipartite graphs to model the problem, harnessing the computational capabilities of deep learning. The conclusive evaluation compares the RAPNet with Murty’s algorithm and the Gibbs sampler, showing accuracy improvements compared to the Gibbs sampler.

Description

Faculties

Fakultät für Ingenieurwissenschaften, Informatik und Psychologie

Institutions

Institut für Mess-, Regel- und Mikrotechnik

Citation

DFG Project uulm

EU Project THU

EVENTS / ReliablE in-Vehicle pErception and decisioN-making in complex environmenTal conditionS / EC / HE / 101069614
PoDIUM / PDI connectivity and cooperation enablers building trust and sustainability for CCAM / EC / HE / 101069547

Other projects THU

AUTOtech.agil / Verbundprojekt MANNHEIM-AUTOtech.agil: Architektur und Technologien zur Orchestrierung automobiltechnischer Agilität / BMBF / 01IS22088W

License

CC BY 4.0 International

Is version of

Has version

Supplement to

Supplemented by

Has erratum

Erratum to

Has Part

Part of

DOI external

DOI external

10.1109/IV55156.2024.10588674

Institutions

Periodical

Degree Program

DFG Project THU

item.page.thu.projectEU

item.page.thu.projectOther

Series

Keywords

Data Association, Graph Neural Network, Zuordnungsproblem, Deep Learning, Tracking (Engineering), Assignment problems (Programming), Deep learning (Machine learning), Neural networks (Computer science), DDC 620 / Engineering & allied operations