Congratulations to Michael Urbiztondo for successfully defending his M.Sc. thesis!

Michael Urbiztondo successfully defended his M.Sc on December 19th, 2025!

Dissertation Title:

Development and Evaluation of an Image-Based Framework for Winter Road Condition Monitoring and Mapping 

Dissertation Description:

His thesis develops an integrated, image-based framework for real-time winter road surface condition (RSC) monitoring and quantitative grip estimation across road networks. The work combines convolutional neural network (CNN) classifiers for automated RSC inference from RWIS and snowplow dashcam imagery with a multi-stage Nested Indicator Kriging (NIK) approach to generate spatially continuous condition maps. In parallel, the thesis evaluates data-efficient deep learning and hybrid regression architectures to predict continuous road-surface grip directly from RWIS imagery under realistic low-data and imbalanced regimes, achieving RMSE values as low as 0.071. The framework is implemented as a cloud-based, operational web platform and is positioned as a transitional foundation for future connected-vehicle hazard sensing and dissemination systems, improving situational awareness, decision support, and winter road safety.

Committee Members:

Dr. Tae J. Kwon (Supervisor), Dr. Stephen Wong (Supervisory Committee Member), Dr. Mustafa Gül (Supervisory Committee Member), and Dr. Vincent McFarlane (Chair)