Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Learning-based Extended Object Tracking Using Hierarchical Truncation Measurement Model with Automotive Radar

Xia, Yuxuan ; Wang, Pu ; Berntorp, Karl Oskar Erik LU ; Svensson, Lennart ; Granstrom, Karl ; Mansour, Hassan ; Boufounos, Petros T. and Orlik, Philip (2021) In IEEE Journal on Selected Topics in Signal Processing 15(4). p.1013-1029
Abstract

This paper presents a data-driven measurement model for extended object tracking (EOT) with automotive radar. Specifically, the spatial distribution of automotive radar measurements is modeled as a hierarchical truncated Gaussian (HTG) with structural geometry parameters, e.g., truncation bounds, orientation, and scaling, that can be learned from the training data. The HTG measurement model provides an adequate resemblance to the spatial distribution of real-world automotive radar measurements. Moreover, large-scale radar datasets can be leveraged to learn the geometry-related model parameters and offload the computationally demanding model parameter estimation from the state update step. The learned HTG measurement model is further... (More)

This paper presents a data-driven measurement model for extended object tracking (EOT) with automotive radar. Specifically, the spatial distribution of automotive radar measurements is modeled as a hierarchical truncated Gaussian (HTG) with structural geometry parameters, e.g., truncation bounds, orientation, and scaling, that can be learned from the training data. The HTG measurement model provides an adequate resemblance to the spatial distribution of real-world automotive radar measurements. Moreover, large-scale radar datasets can be leveraged to learn the geometry-related model parameters and offload the computationally demanding model parameter estimation from the state update step. The learned HTG measurement model is further incorporated into a random matrix based EOT approach with two (multi-sensor) measurement updates: one is based on a factorized Gaussian inverse-Wishart density representation and the other is based on a Rao-Blackwellized particle density representation. The effectiveness of the proposed approaches is verified on both synthetic data and real-world automotive radar nuScenes dataset over 300 trajectories.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Automotive engineering, Automotive radar, autonomous driving, Computational modeling, Density measurement, extended object tracking, nuScenes, Radar, Radar measurements, random matrix, Sensors, Time measurement
in
IEEE Journal on Selected Topics in Signal Processing
volume
15
issue
4
pages
17 pages
publisher
IEEE - Institute of Electrical and Electronics Engineers Inc.
external identifiers
  • scopus:85100862277
ISSN
1932-4553
DOI
10.1109/JSTSP.2021.3058062
language
English
LU publication?
yes
id
dbb84c8b-bb5c-42c5-a458-9ac4598ebb39
date added to LUP
2021-03-04 13:42:21
date last changed
2022-04-27 00:32:52
@article{dbb84c8b-bb5c-42c5-a458-9ac4598ebb39,
  abstract     = {{<p>This paper presents a data-driven measurement model for extended object tracking (EOT) with automotive radar. Specifically, the spatial distribution of automotive radar measurements is modeled as a hierarchical truncated Gaussian (HTG) with structural geometry parameters, e.g., truncation bounds, orientation, and scaling, that can be learned from the training data. The HTG measurement model provides an adequate resemblance to the spatial distribution of real-world automotive radar measurements. Moreover, large-scale radar datasets can be leveraged to learn the geometry-related model parameters and offload the computationally demanding model parameter estimation from the state update step. The learned HTG measurement model is further incorporated into a random matrix based EOT approach with two (multi-sensor) measurement updates: one is based on a factorized Gaussian inverse-Wishart density representation and the other is based on a Rao-Blackwellized particle density representation. The effectiveness of the proposed approaches is verified on both synthetic data and real-world automotive radar nuScenes dataset over 300 trajectories.</p>}},
  author       = {{Xia, Yuxuan and Wang, Pu and Berntorp, Karl Oskar Erik and Svensson, Lennart and Granstrom, Karl and Mansour, Hassan and Boufounos, Petros T. and Orlik, Philip}},
  issn         = {{1932-4553}},
  keywords     = {{Automotive engineering; Automotive radar; autonomous driving; Computational modeling; Density measurement; extended object tracking; nuScenes; Radar; Radar measurements; random matrix; Sensors; Time measurement}},
  language     = {{eng}},
  month        = {{02}},
  number       = {{4}},
  pages        = {{1013--1029}},
  publisher    = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
  series       = {{IEEE Journal on Selected Topics in Signal Processing}},
  title        = {{Learning-based Extended Object Tracking Using Hierarchical Truncation Measurement Model with Automotive Radar}},
  url          = {{http://dx.doi.org/10.1109/JSTSP.2021.3058062}},
  doi          = {{10.1109/JSTSP.2021.3058062}},
  volume       = {{15}},
  year         = {{2021}},
}