Connected Vehicle Data for Risk Analysis and Transportation Performance Evaluation

Abstract

Transportation planning decisions are traditionally based on aggregated metrics, which may obscure crucial temporal and spatial observations of transportation performance. However, the emergence of connected vehicle (CV) data presents an opportunity to obtain new insights into system operations. CV data includes vehicle-specific performance information, reported every few seconds, with metadata such as vehicle speeds, trajectories, and event data such as harsh braking, environmental conditions, and other attributes. This paper showcases the potential of CV data by extending existing methods for transportation performance evaluation and risk analysis. An analysis of CV data is used with a corridor trace analysis for access management, offering support for benefit-cost assessments through intra-hour performance evaluations and analysis of vehicle event data. The dataset used for this investigation comprises 55 billion observations, averaging around 600 million observations per hour, as provided by the Virginia Department of Transportation for demonstration purposes, which yields surprising results. The dataset size requires a robust data architecture that efficiently handles the resource implications outlined in this work. This research demonstrates how transportation professionals and decision-makers can use CV data to make informed decisions in the planning and operations of transportation systems.