FHWA undertook the project, Enhancing Travel Forecasting Practices and Effective Use of Travel Forecasts in Long Range Transportation Planning Through the Application of Artificial Intelligence and Big Data (TOPR HEPP240005PR), in order to, “Investigate the role artificial intelligence (AI) could play in enhancing travel forecasting.” While the project addresses the potential role of AI in travel modeling and forecasting in general, it also included a focused, deep dive on the application of neural networks and AI-DCM models to destination choice.
The Reno, Nevada, metropolitan area was chosen to serve as a case study for the use of AI and Big Data in travel forecasting in part because it offered relatively large and recent datasets on travel patterns including both from a traditional household travel survey as well as big data from connected vehicles. The model incorporated some simple machine learning models similar to those presented in the previous webinar, but the focus of this webinar is on the development of advanced destination choice models incorporating AI and their implementation for the Reno region.
Vince Bernardin, PhD, Vice President at Caliper Corporation, is one of the nation’s foremost experts on the incorporation of both big data and AI in the practice of travel modeling. He was the first to use big data for statewide modeling (2010) and activity-based modeling (2016), and the first to incorporate machine learning/AI methods including boosting (2017) and decision trees (2021). He is serving as Project Manager/Principal Investigator for this project.
Andrew Rohne, is a Senior Transportation Planner at Caliper Corporation. He has over twenty years of experience with over a decade in both the public and private sectors developing, calibrating, and applying both trip and activity-based models. He led the implementation of MWCOG, CMAP, and MetCouncil’s ActivitySim models and also has deep experience with CT-RAMP2 from his time leading OKI’s modeling team. Prior to the current project for FHWA, Andrew had collaborated with Vince using machine learning with big data for commercial vehicle modeling for the Indianapolis MPO.