Transportation Safety and Spatiotemporal Learning
What can incomplete observations tell us about traffic states and safety risks?
We develop statistical and machine learning methods to detect abnormal traffic, forecast evolving risks, and understand interactions among road users. Our emphasis is on sparse observations, transfer across locations, and uncertainty in the available evidence.
Research directions
- Detecting lane-level traffic anomalies from sparse vehicle telematics.
- Modeling the emergence and evolution of crash hotspots across space and time.
- Evaluating risk estimates under changing traffic conditions and sensing coverage.
Selected work
Real-time lane-level abnormal traffic detection on freeways using sparse telematics data
Accident Analysis & Prevention · 2026
Forecasting the Emergence and Evolution of Crash Hotspots: A Unified Deep Learning Framework for Proactive Traffic Safety
Preprint · 2026
A Probabilistic Framework for Estimating the Risk of Pedestrian-Vehicle Conflicts at Intersections
IEEE Transactions on Intelligent Transportation Systems · 2023