Link performance functions for high occupancy vehicle lanes of freeways

Research output: Contribution to journalResearch articleContributedpeer-review

Contributors

  • Xiang Zhang - , University of New South Wales (Author)
  • S. Travis Waller - , University of New South Wales (Author)

Abstract

High Occupancy Vehicle (HOV) lanes are widely used on freeways and play an important role in network design and management. Likewise, link performance functions serve as an essential tool for transport system analysis. This paper aims to support network analysis by providing a tailored link performance function for HOV lanes contiguous with general motor lanes on freeways. Specifically, real traffic data is used for model calibration and evaluation that was assembled from the Performance Measurement System (PeMS) maintained by the California Department of Transportation. Three alternative models for link performance functions of HOV lanes on freeways are developed, which take traffic performance on both HOV lanes and adjacent sets of general motor lanes into consideration. To calibrate the parameters of the models, linear regression is made through stepwise and enter methods and nonlinear regression is carried out using sequential quadratic programming. Statistical analysis together with an evaluation using real traffic data is conducted to evaluate the validity of the proposed models. Our results show that all the three proposed models for contiguous HOV lanes on freeways are statistically significant and perform better in representing real traffic condition with regards to a traditional link performance function, with one specific nonlinear model best supported.

Details

Original languageEnglish
Pages (from-to)657-668
Number of pages12
JournalTransport
Volume33
Issue number3
Publication statusPublished - 1 Jul 2018
Peer-reviewedYes
Externally publishedYes

External IDs

ORCID /0000-0002-2939-2090/work/141543707

Keywords

Keywords

  • Freeway, High occupancy vehicle lane, Link performance function, Regression model, Traffic estimation