Towards Generalizable Drowsiness Monitoring with Physiological Sensors: A Preliminary Study

2025-06-03

AI Summary

In-driver fatigue monitoring, inconsistent definitions and annotations of drowsiness across datasets cause conflicting physiological signal correlations (ECG, EDA, RESP), undermining model generalizability. Method: We systematically analyze four heterogeneous datasets to uncover fatigue-cause-specific physiological response patterns. Leveraging multi-source signal time-frequency feature extraction and cross-dataset consistency modeling, we identify three robust, scenario-invariant biomarkers. Results: These biomarkers are reduced heart rate variability (i.e., increased HRV stability), decreased respiratory amplitude, and lowered EDA baseline. Empirical evaluation shows that a binary logistic regression model built solely on these objective physiological markers significantly outperforms subjectively annotated baselines in sensitivity. Our findings provide theoretical foundations and critical design principles for developing generalizable, privacy-preserving, and subjectivity-free in-vehicle fatigue monitoring systems.

Technology Category

• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Computer Vision: Biometrics, Face, Gesture & Pose
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Security and Privacy: Large-scale security measurements
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Accurately detecting drowsiness is vital to driving safety. Among all measures, physiological-signal-based drowsiness monitoring can be more privacy-preserving than a camera-based approach. However, conflicts exist regarding how physiological metrics are associated with different drowsiness labels across datasets. Thus, we analyzed key features from electrocardiograms (ECG), electrodermal activity (EDA), and respiratory (RESP) signals across four datasets, where different drowsiness inducers (such as fatigue and low arousal) and assessment methods (subjective vs. objective) were used. Binary logistic regression models were built to identify the physiological metrics that are associated with drowsiness. Findings indicate that distinct different drowsiness inducers can lead to different physiological responses, and objective assessments were more sensitive than subjective ones in detecting drowsiness. Further, the increased heart rate stability, reduced respiratory amplitude, and decreased tonic EDA are robustly associated with increased drowsiness. The results enhance understanding of drowsiness detection and can inform future generalizable monitoring designs.

Problem

• Investigates physiological signals for reliable drowsiness detection across datasets
• Examines conflicts in drowsiness metrics due to different inducers and assessments
• Identifies robust physiological markers (ECG, RESP, EDA) linked to drowsiness

Innovation

• Analyzed ECG, EDA, RESP signals across datasets
• Used binary logistic regression for drowsiness metrics
• Identified robust physiological markers for drowsiness