Non-intrusive Driver Fatigue and Stress Monitoring Using Ambient Vibration Sensing

Noh, Hae Young; Zhang, Pei · 2017 · ROSA P / Technologies for Safe and Efficient Transportation. University Transportation Center

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This research addresses the critical safety challenge of monitoring driver state in autonomous vehicles, specifically when control must be handed back to a human operator. As autonomous systems may encounter conditions that blind their sensors, such as severe weather or unexpected road closures, the vehicle must verify that the driver is alert, capable, and physically present before relinquishing control. Existing monitoring solutions often rely on intrusive contact sensors or camera-based systems that are sensitive to lighting and line-of-sight limitations. To overcome these issues, the authors developed a non-intrusive monitoring system using ambient vibration sensing via embedded accelerometers in car seats. The goal was to extract physiological indicators—such as heart rate, breathing, and muscle activity—to infer higher-level states like stress and physical fatigue, despite the high noise levels inherent in a moving vehicle. The methodology combines signal processing with physics-based and data-driven models to isolate small physiological signals from environmental noise. For stress monitoring, the system estimates heart rate and rhythm by analyzing frequency domain features from accelerometers placed near the driver’s heart. It employs an extreme value distribution model to identify and eliminate high-amplitude "outlier" segments caused by vehicle motion or body movement, thereby reducing noise interference. For fatigue monitoring, sensors near the driver’s legs detect minute muscle vibrations (myoelectric signals). The system calculates the Mean Power Frequency (MPF) of these vibrations, operating on the principle that muscle activation frequency shifts to lower values as fatigue increases. This approach allows the system to distinguish between large physical movements and subtle physiological changes, ensuring measurements are taken only when the driver is relatively still. The findings demonstrate that embedded accelerometers can successfully detect key stress and fatigue indicators. Experimental results showed that while raw time-domain heartbeat signals were obscured by environmental noise, frequency analysis revealed distinct harmonic components corresponding to heartbeats, even when signal strength was comparable to noise levels. In fatigue experiments involving isometric exercises, the system observed a clear shift in the power spectral density toward lower frequencies as muscles fatigued. The MPF feature exhibited a decreasing trend over time, confirming that the slope of this metric could effectively estimate the progression of muscle fatigue. These results validate the ability to extract reliable physiological data from high-noise automotive environments. The significance of this work lies in providing a robust, non-intrusive method for assessing driver readiness in autonomous vehicles. By leveraging ambient vibrations, the system avoids the privacy and comfort issues associated with cameras or wearable devices. The ability to accurately infer stress and fatigue levels enables safer handover protocols, ensuring that control is only returned to drivers who are physiologically capable of handling complex driving scenarios. This approach contributes to the broader field of intelligent transportation systems by offering a practical solution for continuous, unobtrusive driver monitoring.

Key finding

Embedded seat accelerometers can reliably extract heart rate and muscle fatigue indicators by isolating harmonic vibration frequencies and detecting downward shifts in mean power frequency despite high environmental noise.

Methodology

lab_experiment

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• drowsiness detection algorithms

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: validation psychometrics