Developing a Platform to Analyze Behavioral Impacts of Connected Automated Vehicles at the National Level

Given that connected autonomous vehicles (CAV) are getting road tested and they may soon become available for public use, significant changes in people’s travel behavior and transportation systems are expected. Nevertheless, their potential impacts on travel behavior have yet to be understood and incorporated into transportation plans. Few studies attempted to answer how the presence of CAVs would affect travel behavior and choices such as time-of-day, trip route, and mode of travel. However, due to lack of experimental data and behavioral surveys on CAV use, the existing literature is limited to developing frameworks, speculation of behavioral changes, or hypothesizing their potential impacts. Specifically, there is yet to be any literature published on large-scale travel demand models that account for CAVs benefits. This study aims to develop a methodological framework that utilizes multiclass data fusion and data transferability techniques, which uses data and models from a smaller geographical area (e.g., Chicago and Detroit) to generate the needed disaggregate data in a larger scale (e.g., national level). This framework is able to perform a comprehensive examination on the impacts of CAV in transportation networks. To achieve this goal, an advanced transportation systems simulation model, POLARIS, which simulates both travel behavior and traffic flow, is used to estimate the potential impacts of CAV technologies at a regional- level. The rich output of CAV scenario analysis in POLARIS framework includes information on person- level and household-level socio-demographic attributes as well as detailed activity-travel patterns. Transferable variables such as total trip rates and travel times are also derived from POLARIS. Following that, the authors use Exhaustive CHAID decision tree models for each transferable variable to cluster people into several homogeneous groups through which various types of lifestyles are captured. The best-fitted statistical distribution for each of the final decision tree clusters is then determined to analyze the specific behavior of members of each cluster toward transferable variables. Finally, using an artificial neural network model, cluster membership rules and travel statistics are transferred to the national level to develop a validated baseline national platform for analyzing connected automated vehicles scenarios. The platform that is capable of transferring travel behavior to national level with high level of accuracy will be utilized to test various policies and scenarios that may affect the use of CAV and their national impacts.

• Supplemental Notes:
  • This paper was sponsored by TRB committee ADB00 Section - Travel Analysis Methods.
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Shabanpour, Ramin
  • Auld, Joshua
  • Mohammadian, Abolfazl (Kouros)
  • Stephens, Thomas S
• Conference:
  • Transportation Research Board 96th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2017-1-8 to 2017-1-12
• Date: 2017

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; References; Tables;
• Pagination: 15p
• Monograph Title: TRB 96th Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Automated vehicle control; Mobile communication systems; Simulation; Travel behavior; Travel demand
• Subject Areas: Data and Information Technology; Highways; Planning and Forecasting;

Filing Info

• Accession Number: 01625855
• Record Type: Publication
• Report/Paper Numbers: 17-06283
• Files: TRIS, TRB, ATRI
• Created Date: Feb 13 2017 2:41PM