A SOLAR-POWERED CONTEXT-AWARE WEARABLE IOT SYSTEM FOR REAL-TIME MICROSLEEP DETECTION

W Ibrahim, Wan Suhaifiza; Abu Bakar, Zahari; Yusuf, Zakariah; Ilham, Nur Iqtiyani · 2026 · Crossref

DOI: 10.35631/ijirev.825034

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 safety risks posed by microsleep episodes—brief, involuntary lapses in consciousness—in critical environments such as driving and industrial operations. Existing detection methods, including electroencephalography (EEG) and computer vision, are limited by invasiveness, high computational demands, lighting sensitivity, and privacy concerns. To overcome these limitations, the authors propose a solar-powered, context-aware wearable Internet of Things (IoT) system designed for real-time, non-invasive microsleep detection with enhanced energy sustainability. The system integrates an ESP32 microcontroller with three primary sensors: an infrared (IR) sensor for eye-closure detection, an accelerometer for head-motion analysis, and a light-dependent resistor (LDR) for ambient light sensing. A key innovation is the brightness-adaptive decision algorithm, which dynamically switches logic based on lighting conditions. In bright environments, the system uses an AND logic, requiring both eye closure and head nodding to trigger an alert, thereby reducing false positives from normal blinks. In dark environments, it employs OR logic, triggering an alert if either condition is met, to maximize sensitivity. Upon detection, the system activates multimodal alerts (visual, auditory, and vibration) and transmits notifications to the Blynk IoT platform. The prototype is powered by a small solar panel and rechargeable battery to support long-term wearability. Experimental validation involved 40 controlled trials performed by a single researcher, comprising 20 trials under bright lighting and 20 under dark lighting. The system achieved an overall detection accuracy of 95%, with 38 correct classifications out of 40 trials. Specifically, the AND logic in bright conditions eliminated false positives but resulted in one false negative, while the OR logic in dark conditions eliminated false negatives but produced one false positive. The system demonstrated an average response time of 1.68 seconds, allowing for intervention within the typical 1–5 second duration of a microsleep episode. Performance metrics included 95% precision, recall, specificity, and F1-score. The findings demonstrate the feasibility of combining hybrid sensor fusion, context-aware adaptive logic, and renewable energy harvesting in a wearable device. The system offers a cost-effective, low-latency alternative to computationally intensive vision-based or invasive EEG systems. While the results are promising, the authors note that validation was limited to a single user in controlled settings. Future work will focus on multi-user testing, real-world field deployment, and the implementation of adaptive thresholding or machine learning techniques to further enhance detection robustness and generalizability.

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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

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

• drowsiness detection algorithms
• microsleep