Crash Data

Crash data analysis aims to understand the causes and consequences of road accidents to improve traffic safety and the design of autonomous vehicles. Current research focuses on developing sophisticated models, including deep learning architectures like convolutional neural networks (CNNs) and transformers, and probabilistic methods like Gaussian Process Regression, to analyze diverse data sources such as police reports, naturalistic driving data, and connected vehicle data. These advancements enable more accurate crash risk prediction, improved safety assessment of autonomous driving systems, and the development of targeted safety interventions, ultimately contributing to a reduction in road accidents and their severity.