Quantum Machine Learning Applied to Human Reliability: EEG Signal Encodings for Drowsiness Detection

The alert status of operators is essential information for the prevention of human error in critical systems. Electroencephalogram (EEG) signals have been widely used to identify brain patterns associated with sleepiness, enabling the development of predictive systems. In this work, we investigate the application of Quantum Machine Learning (QML) algorithms in the classification of EEG signals, focusing on the exploration of different techniques for encoding classical data in quantum circuits. The approach combines quantum neural networks with coding strategies — Angle Encoding, Amplitude Encoding and Pauli Feature Maps — with the aim of analyzing the potential of these strategies to improve sleepiness diagnosis metrics. The results indicate that encodings with greater representation capacity, such as Pauli Feature Maps, achieve better classification performance, while simpler methods, such as Angle Encoding, stand out for their computational efficiency.