Detection and Predictive Analysis of Drowsiness Using Non-contact Doppler Sensor
DOI: 10.31803/tg-20240220092303
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study addresses the critical safety issue of drowsiness-induced accidents, particularly in driving and industrial settings, by proposing a non-contact method for detecting and predicting drowsiness states. While traditional biosignal monitoring methods like electrocardiograms (ECG) and electroencephalograms (EEG) offer high accuracy, they require uncomfortable physical contact with electrodes. The authors aim to overcome this limitation by utilizing a 24 GHz Doppler radar sensor to remotely monitor vital signs, specifically heart rate and heart rate variability (HRV), to identify drowsiness without direct contact. Building on previous work that established over 95% accuracy for heart rate measurement using this technology, this research focuses on developing a predictive model for drowsiness onset. The experimental design involved nine healthy university students who underwent tests to distinguish between wakefulness and induced drowsiness. Drowsiness was induced by sleep deprivation and heavy meals, with the actual onset of drowsiness validated via camera recordings of eye closure. The Doppler sensor, positioned within one meter of the subjects, captured bio-signals which were processed using a custom RRI (R-R interval) algorithm involving bandpass filtering and short-term Fourier transform to isolate heartbeat signals from noise and respiration. The resulting data, including heart rate and HRV metrics, were analyzed using three statistical methods: cross-analysis (chi-square test), logistic regression, and panel logistic regression. The panel analysis specifically incorporated time lags to assess the predictive capability of the sensor data relative to the actual moment of drowsiness. The results demonstrated a statistically significant association between the calculated drowsiness stage and the actual drowsiness state, with a p-value of less than 0.001. Cross-analysis revealed that drowsiness typically began at stage 3 and was confirmed in over 95% of cases at stages 4 and 5. Logistic regression indicated that for every one-level increase in the drowsiness stage, the likelihood of being in a drowsy state increased by approximately 5.554 times. Crucially, the panel logistic regression analysis identified an optimal prediction window; the strongest statistical relationship occurred with a time lag of 20 to 30 seconds, where the odds ratio peaked at 2.105. This indicates that the sensor data can predict the onset of drowsiness approximately 20–30 seconds before it is consciously recognized or visually confirmed. The significance of this research lies in its ability to provide an early warning system for drowsiness using non-invasive technology. By accurately predicting the timing of drowsiness onset with a 20–30 second lead time, the system offers a practical solution for preventing accidents caused by fatigued drivers or workers. This approach eliminates the discomfort and inconvenience of wearable sensors while maintaining high reliability. The authors conclude that this technology has strong potential for application in automotive safety systems and medical monitoring, suggesting future research should expand the sample size and integrate artificial intelligence for further data refinement.
Key finding
Non-contact Doppler radar can detect driver drowsiness and predict drowsiness onset with lead time, validated against eye-closure ground truth.
Methodology
lab_experiment
Sample size: 9
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 discover_doaj on 2026-05-03 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-03 |
| archive | success | — | — | — | 1 | 2026-05-03 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | crossref | — | — | 3 | 2026-06-04 |
| promote | success | — | — | — | 1 | 2026-05-03 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 17 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
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.
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