Research directions that connect data, models, and transportation safety decisions.
Interpretable crash-risk evidence through causal mediation, random-parameter models, Bayesian structure, and spatial effects.
First-person near-miss and crash event understanding with traffic video, LLM/VLM event parsing, and counterfactual safety analysis.
Real-time freeway and arterial safety assessment with connected vehicle trajectories, risky-driving events, video, and sequence learning.
AI-driven measures of effectiveness for speed management, freeway operations, and safety-mobility coupling.