A novel pull-out semi-dry electrode with rapid conductive liquid replenishment for EEG acquisition in real driving environment

To address the issues of convenience, comfort, and noise resistance in the detection of driving fatigue in real driving environments, this study developed a novel pull-out semi-dry electrode and constructed a driving fatigue recognition model using a one-dimensional deep residual shrinkage network (1D-DRSN). The novel electrode features a spring reset structure for rapid conductive liquid replenishment, combined with silver-plated electrode cores and PU sponge, ensuring high signal quality and comfortable wear. The 1D-DRSN model integrates residual connections, attention mechanisms, and soft threshold functions to effectively reduce noise interference, demonstrating high accuracy and robustness. Results show that the novel electrode can achieve up to 10 hours of effective electroencephalography (EEG) signal acquisition, and the 1D-DRSN model achieves an average classification accuracy rate of 99.65 % in driving fatigue detection tasks, maintaining excellent performance even in noisy environments. This study provides an efficient and reliable solution for signal acquisition and detection in the field of driving fatigue detection.