Automated Vehicle Modeling Peer Exchange

1 post / 0 new
Automated Vehicle Modeling Peer Exchange

The Travel Model Improvement Program has released a report on a peer exchange conducted in June 2019 to discuss Automated Vehicle Modeling by Metropolitan Planning Organizations.  The report may be accessed at this link:

With increasing interest in AVs and the potential impacts they stand to have on the Nation's transportation network, many transportation agencies, including Metropolitan Planning Organizations (MPOs), are actively exploring ways to address AV modeling considerations in the long-range transportation planning process. The peer exchange convened representatives of five MPOs to discuss their approaches to AV modeling:

Maricopa Association of Governments (MAG) (Phoenix, AZ)

Mid-America Regional Council (MARC) (Kansas City, MO)

North Central Texas Council of Governments (NCTCOG) (Arlington, TX)

Sacramento Area Council of Governments (SACOG) (Sacramento, CA)

San Diego Association of Governments
(SANDAG) (San Diego, CA)

Key findings of the peer exchange include the following:

  •        There is tremendous uncertainty about the impacts AVs will have on future transportation systems. Behavioral travel
    models are among the few tools that may provide useful mechanisms for conceptualization of the effect of these technologies.

  •        Using models can allow for more informed decisionmaking, recognizing that models provide insights and not answers. Every model has underlying assumptions and data. Models are a way to quantify the future and can provide a consistent set of metrics to explore scenarios reflecting various assumptions (e.g., land use, socioeconomic data, etc.).

  •       Scenario testing can make the most of AV modeling compared to other more specific forecasting techniques due to the approach's flexibility in exploring outcomes of multiple, varied futures.

  •        Organizing estimates of "uncertainty" factors in the modeling chain by category of assumptions could help MPO modelers better understand how the factors relate and potentially impact each other.

  •        Presenting model results and related information should include consequential assumptions about the model and external inputs. This method of communication provides a meaningful context for interpretation of the model results. In addition, it is important to keep the target audience in mind so that the model's underlying assumptions and data can be communicated effectively.

  • Data interpretation is the foundation of travel modeling and is a significant skillset needed for modeling professionals. New data, new applications, and new data management approaches are emerging from AVs and other new technologies.