Multimodal Fatigue Detection and Management System

Research Scientist | Technical Project Lead

Leading the development of a multi-modal fatigue detection system to better understand and predict human states of fatigue and stress in individuals with high-stress jobs such as emergency medical service professionals.

System built
Multimodal fatigue detection pipeline
Wearables + self-reports + context signals
Core goal
Predict and track fatigue under real operations
Longitudinal field monitoring in high-stress roles
Outcome
Actionable, human-centered feedback
Signals translated into operationally usable insights

The problem

Fatigue is cumulative, context-dependent, and hard to detect early
Self-reports are noisy and delayed relative to physiological changes
Operational environments add confounds like stress and workload

Data & sensing

Wearable physiology for continuous monitoring
Longitudinal data collection across real shifts
Context signals to interpret operational demands

Modeling approach

Feature engineering from physiological time series
Prediction and forecasting of fatigue and stress states
Validation with human-centered evaluation metrics

Why it matters

Supports safer decision-making in high-stakes work
Enables early detection and targeted intervention
Turns raw sensor streams into deployable insights

Technology stack

Wearable Physiology Time-Series Feature Engineering Machine Learning Longitudinal Study Design Human Factors Evaluation Operational Analytics

Conference Proceedings