Managing the Impacts of Different CV/AV Penetration Rates on Recurrent Freeway Congestion From the Perspective of Traffic Management: A Case Study of MD-100

Project Abstract

In the near future, responsible highway agencies will need to effectively coordinate emerging AV flows while contending with daily recurrent congestion. This study presents a systematic procedure for understanding how AV flows impact traffic under different AV behavioral mechanisms (i.e., car-following and lane-changing), penetration rates, and volume levels. Using a congested segment of the MD-100 highway to illustrate the proposed procedure, our research results indicate that the presence of AV flows, depending on their adopted behavioral mechanisms, significantly impacts (either positively or negatively) the overall traffic conditions. These impacts, varying with AV penetration rate and volumes, will be experienced indiscriminately by AV and non-AV vehicles. The study has further conducted extensive simulation experiments using the MD-100 network under various AV penetration rates and behavioral mechanisms by modeling the range of the behavioral mechanisms likely adopted by the AV-flows with 135 sets of car-following and lane changing parameters. The collected measures of effectiveness (MOEs) from the experimental results clearly show that at each AV penetration level, there exists a set of optimal behavioral mechanisms for the AV flows to coordinate with non-AV flows to best use roadway capacity and minimize congestion. Since such behavioral mechanisms vary with the AV penetration rate and the congestion level on different segments of the freeway, it justifies the need for a responsible highway agency to develop effective guidelines so that they can coordinate with the AV flows via the V2I infrastructure.

Read the full report.

Outputs and Outcomes

A responsible highway agency can follow our proposed method to develop operational guidelines that will enable the traffic operators to properly coordinate with AV flows to make the best use of the roadway capacity and avoid any potential negative impacts of AVs.

Universities and Sponsoring Organizations Involved

The University of Maryland, College Park

U.S. Department of Transportation Office of the Secretary-Research

Principal Investigator

Dr. Gang-Len Chang, Email: gang@umd.edu

Funding Sources and Amounts

USDOT: $100,000; University of Maryland: $50,000 (match)

Start Date

July 1, 2018

Completion Date

November 2019

Expected Research Outcomes

Research findings from all project tasks will be organized into a friendly, rule-based expert system that can assist traffic agencies responsible for daily freeway traffic monitoring/management in communicating with CV flows with respect to the set of behavioral parameters they should be programmed to on different segments of the target freeway, based on the detected traffic conditions.

Keywords

Autonomous vehicles, automated driving, traffic flow management, heterogeneous traffic