Evaluating System Impacts of Automated Vehicles: A Multiple-Resolution Modeling Approach


In this SBIR project, we propose an Automated Vehicle System Impacts Evaluation Module, AVSIEM, for both transportation and traffic operation analysis using a two-step multi-resolution modeling approach.  The approach hinges on the development of a Capacity Adjustment Factor (CAF) for automated vehicles, similar to the heavy vehicles adjustment factor used in highway capacity analysis.  CAF will be linked to input variables such as roadway facility types, traffic demand levels, and market penetration rates of automated vehicles.  CAF will be derived from a microsimulation environment, which involves the development of an integrated car-following model for both human drivers and automated vehicles, model calibration using vehicle trajectory data from real-world, and implementation of the model as a plugin in microsimulation.  The derived CAF will then be introduced into a macroscopic regional network for adjusting capacity and the impacts of automated vehicles can be analyzed through additional traffic assignment runs.


Background and Objectives

The rapid development of automated vehicles has attracted a lot of attentions from the public in recent years.  Current studies on automated vehicles mainly focus on microscopic simulations with simple network topologies and driver behaviors, and few has considered to incorporate automated vehicles into macroscopic travel demand models for the analysis in a regional network.  The proposed AVSIEM connects the qualitative analysis of automated vehicles using microsimulation with macroscopic-level system impact evaluation of the automated vehicles on transportation networks.  With the goal to successfully demonstrate the proof of concept of AVSIEM, this project will design and develop the CAF to connect microsimulation analysis results with regional travel demand forecasting models to evaluate the system impacts of automated vehicles.


Accomplished Tasks

Phase I of this project focused on the development of a prototype AVSIEM through a single lane network.  The tasks include:

  • Comprehensive review of existing car-following models and regional transportation models
  • Development of an integrated car-following model for both human drivers and automated vehicles
  • Implementation of the integrated car-following model as a plugin in microsimulation
  • Calibration of the integrated car-following using vehicle trajectory data from the real-world
  • Prototype roadway networks to evaluate the single-lane link performance with the application of automated vehicles for different road geometries and signal control settings
  • Development of the CAF to address the impacts of automated vehicles using microsimulation
  • Commercialization analysis to show potential applications of AVSIEM


Sponsor: USDOT/RITA/Volpe Center

Project Time Period: Sep 2015 - Feb 2016 (Phase I)