Research
Enhancing Traffic Safety Modeling with CV Driving Behaviors
The proposed traffic safety modeling framework identifies risky driving behaviors (e.g., hard braking, hard turning, lane-changing) from CV trajectories and integrates these features into macro crash frequency modeling and micro real-time crash predictions, achieving superior prediction performances and better interpretability compared to traditional methods.
CV-Data-Based Network Traffic Operation and Prediction
Leveraging emerging CV data, we develop traffic state evaluation methods to extract comprehensive traffic operation indicators (e.g., travel speed, delay, average queue length, Level of Service) for both freeway and urban arterial networks. Spatiotemporal DL models have been used fot network-scale traffic predictions of traffic speed, delay, queue length, and other critical metrics.
Innovative Analytic Methods for Crash Research
My collaborators and I develop novel crash analytic methods to address complex heterogeneity and spatiotemporal dependencies in crash frequency and severity modeling. These approaches encompass both advanced statistical methods and spatio-temporal machine learning/deep learning methods.