Using link travel time co variance information to predict dynamic journey times in stochastic road networks

H. W. Ho, William H.K. Lam, Mei Lam Tam

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

Abstract

Journey time prediction is a crucial component in advanced traveler information systems for helping travelers in making their travel decisions. This paper investigates the journey time prediction problem in road network with stochastic journey times and link flows. The proposed prediction framework consists of two sub-modules. The first one is a reliability-based dynamic traffic assignment model to establish a database for the historical traffic conditions, while the other sub-module, which is a multi-level k-NN model for predicting journey times based on the historical records in the database. A Sioux Falls road network example is used to demonstrate the accuracy, efficiency and robustness of the proposed framework for the journey time prediction problem in stochastic network with uncertainties.

Original languageEnglish
Title of host publicationTransport and Society - Proceeding of the 22nd International Conference of Hong Kong Society for Transportation Studies, HKSTS 2017
EditorsAnthony Chen, Tony N.N. Sze
Pages159-166
Number of pages8
ISBN (Electronic)9789881581464
Publication statusPublished - 2017
Externally publishedYes
Event22nd International Conference of Hong Kong Society for Transportation Studies: Transport and Society, HKSTS 2017 - Hong Kong, Hong Kong
Duration: 9 Dec 201711 Dec 2017

Publication series

NameTransport and Society - Proceeding of the 22nd International Conference of Hong Kong Society for Transportation Studies, HKSTS 2017

Conference

Conference22nd International Conference of Hong Kong Society for Transportation Studies: Transport and Society, HKSTS 2017
Country/TerritoryHong Kong
CityHong Kong
Period9/12/1711/12/17

Keywords

  • Dynamic traffic assignment
  • Effective path journey time
  • Journey time prediction
  • K-nearest neighborhood
  • Travel time covariance

Fingerprint

Dive into the research topics of 'Using link travel time co variance information to predict dynamic journey times in stochastic road networks'. Together they form a unique fingerprint.

Cite this