Evangelos I. Kaisar, Ph.D. (PI)
Professor, Department of Civil, Environmental and Geomatics Engineering
Florida Atlantic University
ekaisar@fau.edu

Praveen Edara, Ph.D.
Professor, Department of Civil and Environmental Engineering
University of Missouri
edarap@missouri.edu

 

Proposal Summary and Objectives

Freight movements are expected to increase by approximately 42 percent by 2040. Trucks are expected to show the largest increase in flows by 2040 across all freight modes. However, the ability for transportation agencies to understand and adequately plan for increased truck movement and related impacts is still limited due to a lack of data on truck travel patterns. The main sources of truck data are truck surveys and truck counts collected by infrastructure-embedded sensors. Although surveys provide detailed information (i.e., truck type, Origin-Destination, weight, and vehicle miles traveled) useful for understanding truck activity patterns, they cannot be utilized to quantify truck activity at the geographical level due to low response rates. Connected vehicle (CV) data availability has been exploding in recent years. This is as a result of the advent of OEMs, Telematics platforms, and other in-vehicle technologies, that are able to continuously stream high-resolution, reliable and accurate vehicle data. The goal of this study is to explore new opportunities for freight activity monitoring by integrating this rich dataset with existing public and private freight datasets to quantify truck activity across the State. In Phase 1 of the project, a spatio-temporal conflation framework that enables seamless integration of three key freight data sources including weigh-in-motion (WIM), freight facility, and traffic flow data was developed. A massively parallel database was then designed to store the integrated data on a cluster of servers enabled with Graphical Processing Units (GPUs). While emerging CV data could provide valuable insights into truck activity patterns, the sheer volume and speed of this data can be overwhelming and challenging to mine with conventional data processing pipelines. The need for frameworks that are able to leverage recent advances in big data and cloud computing to integrate, analyze and interactively visualize freight activity patterns from these new technologies is therefore crucial. The objectives of this study are therefore to:

1. Develop a set of routines for integrating connected vehicle data with traditional freight data sources to detect and analyze freight activity patterns on a large scale,
2. Leverage high-performance computing to develop a scalable database for storing and retrieving integrated datasets,
3. Deploy an interactive, web-based data visualization platform for exploring freight activity patterns.

Funding Amount:
Status: Active
Duration: Sep 1, 2021 - Jul 30, 2022