Skip to main navigation Skip to search Skip to main content

TRAFFIC PARAMETERS ESTIMATION WITH AN ITERATIVE PARTIAL BACKPROPAGATION MAXIMUM LIKELIHOOD ESTIMATION (IPB-MLE) FRAMEWORK

  • Hong Kong University of Science and Technology

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

Abstract

This study proposes a numerical framework for traffic parameters estimation, namely Iterative Partial Backpropagation Maximum Likelihood Estimation (IPB-MLE). This framework utilizes analytical continuations to avoid the discrete nature of the Poisson distribution, hence providing more accurate convergences. Backpropagation introduced in this framework contributes flexibility and possibilities of introducing customizations to the problem formulation. In the experiment, the IPB-MLE framework yields better accuracy and robustness in satisfactory computational times. Estimation takes on average 2 seconds for 10% penetration rate, and the computational complexity increases only linearly with the amount of data and number of variables. Therefore, the IPB-MLE framework offers a great potential for modeling more complex but realistic situations.

Original languageEnglish
Title of host publicationProceedings of the 25th International Conference of Hong Kong Society for Transportation Studies, HKSTS 2021
Subtitle of host publicationSustainable Mobility
EditorsRyan C.P. Wong, Jiangping Zhou, W.Y. Szeto
PublisherHong Kong Society for Transportation Studies Limited
Pages375-382
Number of pages8
ISBN (Electronic)9789881581495
StatePublished - 2021
Externally publishedYes
Event25th International Conference of Hong Kong Society for Transportation Studies: Sustainable Mobility, HKSTS 2021 - Hong Kong, Hong Kong
Duration: 9 Dec 202110 Dec 2021

Publication series

NameProceedings of the 25th International Conference of Hong Kong Society for Transportation Studies, HKSTS 2021: Sustainable Mobility

Conference

Conference25th International Conference of Hong Kong Society for Transportation Studies: Sustainable Mobility, HKSTS 2021
Country/TerritoryHong Kong
CityHong Kong
Period9/12/2110/12/21

Keywords

  • Analytical continuation
  • Backpropagation
  • Flow estimation
  • Intersection
  • Maximum likelihood estimation (MLE)

Fingerprint

Dive into the research topics of 'TRAFFIC PARAMETERS ESTIMATION WITH AN ITERATIVE PARTIAL BACKPROPAGATION MAXIMUM LIKELIHOOD ESTIMATION (IPB-MLE) FRAMEWORK'. Together they form a unique fingerprint.

Cite this