Citation

Rabbi, Riadul Islam and Em, Poh Ping and Hossen, Jakir (2025) Heart Rate Variability-Based Driver Drowsiness Detection Using Recurrent Neural Networks. In: 2025 Multimedia University Engineering Conference, MECON 2025, 21 July 2025 - 23 July 2025, Cyberjaya, Malaysia.

Text 3.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Driver drowsiness contributes significantly to road accidents, necessitating the development of effective real-time monitoring systems. Traditional drowsiness detection systems utilize behavioral or vehicular parameters but consistently lack real-time monitoring of actual physiological indicators. Heart Rate Variability (HRV) exhibits a great potential physiological marker to detect drowsiness because it shows how the autonomic nervous system operates while revealing changes in fatigue and alertness. The research analyses the application of real-time HRV data through different recurring neural network (RNN) architectural designs for driver drowsiness detection levels. The objective aims to identify driver drowsiness states, which include alert, early drowsiness, and severe drowsiness. The methodology incorporates three RNN architectures—Simple RNN, Bi-directional RNN, and Deep RNN—alongside a preprocessing approach using a band-pass filter on heart rate data. Additionally, the Lomb-Scargle Periodogram method is applied in the frequency domain to extract features. Experimental findings show that a simple RNN achieves 86% of the accuracy through Adam optimizer training, and a bidirectional RNN reaches 87.00 % accuracy with the stochastic gradient descent (SGD) optimizer. Deep RNN demonstrated the most promising performance, achieving a classification accuracy of 95.68%. These findings reveal the potential of deep learning-based methods to detect drowsiness in real time through physiological signals. The study underscores that deep RNNs are highly effective in extracting temporal patterns from HRV data, enabling advances in driver monitoring systems with enhanced road safety benefits.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	driver drowsiness, RNNs, HRV, physiological signals, wearable devices
Subjects:	T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL1-484 Motor vehicles. Cycles
Divisions:	Faculty of Engineering and Technology (FET)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	18 Mar 2026 08:14
Last Modified:	19 Mar 2026 01:53
URII:	http://shdl.mmu.edu.my/id/eprint/15583

Downloads

Edit (login required)