Random-Matrix Based Extended Object Tracking Using Multiple Sensors with Different Measurement Matrices

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Random-matrix based extended object tracking (EOT) has received much attention for the simple yet effective scheme. Multiple sensors can bring more information due to the different perspectives and characteristics especially for EOT. However, in practice, multiple sensors have different measurement matrices, which makes EOT even more challenging. This paper develops a variational Bayesian approach to multi-sensor EOT (VB-MS-EOT) based on random matrix. First, the system model of MS-EOT with different measurement matrices is developed. This model also considers the differently distorted observations and various measurement numbers of multiple sensors. To deal with the different measurement matrices and distortions, VB-MS-EOT is proposed, where the kinematic state and extension are estimated iteratively with a simple form. Compared with the traditional EOT based on random matrix using multiple sensors with identical measurement matrix, the effectiveness of VB-MS-EOT is illustrated by simulated data.

Original languageEnglish
Title of host publicationICARCE 2023 - 2023 2nd International Conference on Automation, Robotics and Computer Engineering
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350308341
DOIs
StatePublished - 2023
Event2nd International Conference on Automation, Robotics and Computer Engineering, ICARCE 2023 - Virtual, Online, China
Duration: 14 Dec 202316 Dec 2023

Publication series

NameICARCE 2023 - 2023 2nd International Conference on Automation, Robotics and Computer Engineering

Conference

Conference2nd International Conference on Automation, Robotics and Computer Engineering, ICARCE 2023
Country/TerritoryChina
CityVirtual, Online
Period14/12/2316/12/23

Keywords

  • different measurement matrices
  • extended object tracking
  • multisensor measurements
  • random matrix

Fingerprint

Dive into the research topics of 'Random-Matrix Based Extended Object Tracking Using Multiple Sensors with Different Measurement Matrices'. Together they form a unique fingerprint.

Cite this