Modeling Cooperative Driving Behavior in Freeway Merges

Merging locations are major sources of freeway bottlenecks. Microscopic simulation tools are receiving increased popularity in analyzing these bottlenecks and designing optimum geometric configurations and control strategies. In congested situations, acceptable gaps for merging are often not available and freeway mainline drivers often cooperate with the on-ramp drivers and create gaps for the merge. This is usually done either by decelerating or by changing to an inner freeway lane. Existing approaches do not capture these phenomena and as a result often over-predict congestion. As part of Next Generation SIMulation (NGSIM) project of the Federal Highway Agency (FHWA), we have previously developed a model for the decisions of the merging driver that captures normal, courtesy and forced merges of the driver in a single decision framework. In the current project, we propose to extend the model to include the driving decisions of the freeway mainline drivers in the merging area. The developed model will be an integrated driving behavior model for lane-changing and acceleration decisions of freeway mainline drivers approaching a merging on-ramp that explicitly captures courtesy and cooperative actions of the driver. Parameters of the model will be estimated with vehicle trajectory data collected by NGSIM at Interstate-80, California during congested periods. The models will be implemented in microscopic traffic simulator MITSIMLab to improve its performance in congested merge locations. The enhanced MITSIMLab will be open-sourced and can serve as an excellent tool for transport professionals to test different congestion management mechanisms. Merging locations are major sources of freeway bottlenecks and are therefore important for freeway operations analysis. Microscopic simulation tools have been successfully used in analysis of merging bottlenecks and to design optimum geometric configurations and control strategies for such locations. In congested situations, acceptable gaps for merging are often not available and freeway mainline drivers often cooperate with the on-ramp drivers and create gaps for the merge. This is usually done either by decelerating or by changing to an inner freeway lane. Also, in some cases, the merging driver may become impatient and decide to force in compelling the lag driver in the freeway to decelerate. The lane-changing and acceleration decisions of the freeway mainline driver are therefore not only based on his present situation, but also influenced by the anticipated intention of the merging driver (e.g. whether or not he is executing a forced merge). After evaluation of the state-of-the art merging models, merging models have been identified as a prioritized Core Algorithm by NGSIM (Cambridge Systematics 2003). As part of NGSIM, we have developed a combined model for the decisions of the merging driver (Choudhury et al. 2006). In this model, the decisions of the freeway mainline driver are treated as external variables. The operational validation study of that model showed improvements in the simulation capability compared to the benchmark model. But the improvements can be more significant by detailed modeling of the decisions of the freeway mainline drivers. In this project, we will make the research complete by modeling the behavior of the freeway mainline driver in a congested situation. The developed model will be an integrated driving behavior model for lane-changing and acceleration decisions of a freeway mainline driver in the vicinity of an on-ramp. The model will explicitly capture courtesy and cooperative actions of the driver and take into account the anticipated actions of the merging driver in the decision framework. The model will be estimated with trajectory data and account for the unobserved heterogeneity among the drivers. The validity of the cooperative lane-changing model will be demonstrated by the improvements in the simulation capability of the microscopic traffic simulator MITSIMLab using aggregate traffic data.

    Language

    • English

    Project

    • Status: Completed
    • Funding: $130000.00
    • Contract Numbers:

      DTRT07-G-0001

      MITR20-1

    • Sponsor Organizations:

      Research and Innovative Technology Administration

      Department of Transportation
      1200 New Jersey Avneue, SE
      Washington, DC  United States  20590
    • Performing Organizations:

      Massachusetts Institute of Technology, Cambridge

      Public Affiliates Program, Center for Transportation Studies
      77 Massachusetts Avenue, Room 1-153
      Cambridge, MA  United States  02139-4307
    • Principal Investigators:

      Choudhury, Charisma

    • Start Date: 20080901
    • Expected Completion Date: 0
    • Actual Completion Date: 20090831
    • Source Data: RiP Project 19933

    Subject/Index Terms

    Filing Info

    • Accession Number: 01480785
    • Record Type: Research project
    • Source Agency: New England University Transportation Center
    • Contract Numbers: DTRT07-G-0001, MITR20-1
    • Files: UTC, RiP
    • Created Date: May 7 2013 1:02AM