Advanced Behavioral Analysis of High Resolution Mobility Data

Activity-based models (ABMs) are now well established as the state-of-art in modeling travel demand and supporting decision-making. One challenge to ABMs is their need for detailed, high resolution activity data. Recent developments in smartphone-based travel surveys have helped overcome this challenge, but these new data complicate the modelling. Identifying the correct model structure and specification is difficult. Inter- and intra-person heterogeneity exists also across choice dimensions and the boundaries between primary and non-primary activities is becoming more blurry. This project explores a data-driven non-parametric approach for daily activity pattern generation without behavioral assumptions: Variational Auto-encoder (VAE). This research aims to make four contributions to transportation modeling: deriving realistic activity patterns that resemble the level of heterogeneity in real data; providing an interpretable & explorative tool; allowing for rapid model development and estimation; and being useful for forecasting & policy analysis. In the end, the research team aims to facilitate ABM modeling, making the models more richly reflect real-world conditions while also making them more computationally tractable and, ultimately, more easily operationalized in forecasting and planning practice.

Language

  • English

Project

  • Status: Active
  • Funding: $100000
  • Contract Numbers:

    DTRT13-G-UTC13

    Project MITR25-56

  • Sponsor Organizations:

    Office of the Assistant Secretary for Research and Technology

    University Transportation Centers Program
    Department of Transportation
    Washington, DC  United States  20590
  • Managing Organizations:

    New England University Transportation Center

    Massachusetts Institute of Technology
    77 Massachusetts Avenue, Room 40-279
    Cambridge, MA  United States  01239
  • Project Managers:

    Coughlin, Joseph

  • Performing Organizations:

    Massachusetts Institute of Technology, Cambridge

    77 Massachusetts Avenue
    Cambridge, MA  United States  02139
  • Principal Investigators:

    Zegras, Christopher

  • Start Date: 20180901
  • Expected Completion Date: 20190630
  • Actual Completion Date: 0
  • USDOT Program: University Transportation Centers Program

Subject/Index Terms

Filing Info

  • Accession Number: 01707484
  • Record Type: Research project
  • Source Agency: New England University Transportation Center
  • Contract Numbers: DTRT13-G-UTC13, Project MITR25-56
  • Files: UTC, RiP
  • Created Date: Jun 3 2019 1:34PM