A Cascaded Deep Forest Framework for Robust Driver Fatigue Detection using Forehead Electroencephalography

Performance decrement due to fatigue is a leading contributor to traffic accidents and fatalities.Electroencephalogram (EEG) is widely accepted as a reliable physiological indicator of cognitive state, though the application of EEG-based systems in driver monitoring is often limited by the need for multichannel headsets.Forehead EEG, in particular, has emerged as a promising candidate for early detection due to the rise of portable wearable devices.In this work, we propose a robust and efficient method for driver fatigue detection using forehead EEG signals.The approach employs a cascaded deep forest (CDF) framework, incorporating wavelet log-energy entropy and high-order component statistics to extract meaningful features from low-channel EEG signals.A comprehensive labelling protocol was conducted across 26 subjects to validate the method.The experimental results demonstrated a significant improvement in performance, achieving an average accuracy of 95.1%, which outperformed previous studies.Furthermore, the energy characteristics of small-scale oscillations in brain signals across different frequency bands, along with the application of higherorder statistics in the reconstructed phase space, were validated for computational efficiency.This study presented a new framework of using frontal EEG based on a cascade structure to construct a landing fatigue detection method.It could also provide a promising approach for biomedical signal processing in low-channel systems.