Toward Vehicle-Agnostic Driving Signatures for Cognitive Impairment Prediction from Naturalistic Driving Data

DRIVES Project, Washington University in St. Louis — Master's Thesis (Thesis Supervisor & Lab PI: Dr. Ganesh Babulal; Thesis Advisor & Committee Chair: Dr. Alvitta Ottley) — Aug 2025 – May 2026

St. Louis, Missouri, USA

Research

Responsibilities & Contributions

  • Built an end-to-end pipeline processing 26,968 participant-weeks of naturalistic driving data from 304 older adults, deriving weekly driving features and integrating demographic covariates to predict Clinical Dementia Rating (CDR) status.
  • Benchmarked six model families (logistic regression, random forest, XGBoost, MLP, DANN, GRU-DANN) under leave-one-participant-out cross-validation; GRU-DANN achieved the highest participant-level ROC AUC (0.599) and balanced accuracy (0.584).
  • Defended April 2026; published as no. 1341 in the McKelvey School of Engineering Graduate Student Theses & Dissertations.

Links

← Back to all experience