Proceedings of the

The 33rd European Safety and Reliability Conference (ESREL 2023)
3 – 8 September 2023, Southampton, UK

Quantum Machine Learning for Drowsiness Detection with EEG Signals

Lavínia Maria Mendes Araújo1,a, Plínio Márcio da Silva Ramosa1, Isis Didier Lins1,b, Caio Bezerra Souto Maior1,2,c, Rafael Chaves Souto Araújo3, Andre Juan Ferreira Martins de Moraes4,e, Askery Alexandre Canabarro4,f, Márcio José das Chagas Moura1,d and Enrique López Droguett5

1Center for Risk Analysis and Environmental Modeling, Department of Industrial Engineering, Federal University of Pernambuco, Recife, Brazil.

2Technology Center, Universidade Federal de Pernambuco, Caruaru, Brazil /EADDRESS/
3International Institute of Physics, Federal University of Rio Grande do Norte, Natal, Brazil.

4Department of Physics, Federal University of Alagoas, Arapiraca, Brazil.

5University of California, Los Angeles, United States of America.


Human reliability is increasingly important in accident prevention, and monitoring biological parameters can help detect patterns indicating behaviors that may lead to accidents. Electroencephalogram (EEG) data has been used to identify drowsiness, a major cause of fatigue in machine operators in the oil and gas industry. While classic machine learning methods like Multilayer Perceptron (MLP) have been used with EEG data, quantum computing has shown promise in solving complex problems efficiently. Variational Quantum Algorithms are one example of quantum concepts applied to classical structures for data training. This study aims to classify operator drowsiness using Quantum Machine Learning (QML) models. EEG signals are preprocessed to extract relevant features such as Higuchi Fractal Dimension, Complexity, and Mobility, as well as statistical features. QML models are trained with various quantum circuit layers, rotation, and entanglement gates. Results will be compared with classical MLP models. This work contributes to exploring the context of drowsiness in QML, which has not been extensively studied in the literature. It serves as a proof of concept that QML models are suitable for this type of data and can be further improved as Quantum Computing continues to evolve.

Keywords: EEG. Quantum machine learning, Drowsiness detection, Diagnosis, Variational quantum algorithm.

Download PDF