We develop methods to understand transportation systems from incomplete observations, build AI models grounded in transportation knowledge, and connect sensing with useful digital representations.
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.
We study how vision-language models interpret roadway video, connect vehicle and infrastructure views, and reason about dynamic events. We build benchmarks and learning methods to evaluate what models observe, infer, and generalize in transportation settings.
We connect roadway sensing, learned representations, and simulation to develop transportation digital twins. Our work examines lane geometry and behavior, physical–virtual integration, and when cooperation between vehicles and infrastructure is useful.
Developing lightweight vision-language models to interpret precipitation, visibility, and road surface conditions from roadside cameras. Planned outputs include a weather visual question-answering dataset and evaluation on edge computing platforms.
Wyoming NASA EPSCoR Faculty Research Award
Linking vehicle observations across sparse roadside cameras through vehicle re-identification and transportation-aware spatiotemporal reasoning, with the aim of supporting travel-time estimation and anomaly detection in rural corridors.
UW CEPS Engineering Initiative seed grant program
The proposed center focuses on multimodal sensing, AI under sparse data and limited connectivity, and corridor-scale digital twins for rural transportation safety, operations, and workforce development.
University of Wyoming Science Institute
Integrated travel-time, traffic, crash, incident, lane-closure, and weather data to measure recurring and nonrecurring highway delay and examine associated contributing factors.
Wisconsin Department of Transportation
