This study presents a novel driver authentication system utilizing electrocardiogram (ECG) signals collected through dry electrodes embedded in the steering wheel. Traditional biometric authentication methods are sensitive to environmental changes and vulnerable to replication, but this study addresses these issues by leveraging the unique characteristics and forgery resistance of ECG signals. The proposed system is designed using autocorrelation profiles (ACPs) and a convolutional neural network and is optimized for real-time processing even in constrained hardware environments. Additionally, advanced signal processing algorithms were applied to refine the ECG data and minimize noise in driving environments. The system’s performance was evaluated using a public dataset of 154 participants and a real-world dataset of 10 participants, achieving F1-Scores of 96.8% and 96.02%, respectively. Furthermore, an ablation study was conducted to analyze the importance of components such as ACPs, normalization, and filtering. When all components were removed, the F1-Score decreased to 60.1%, demonstrating the critical role of each component. These findings highlight the potential of the proposed system to deliver high accuracy and efficiency not only in vehicle environments but also in various security applications.

Accurate sleep stage classification is essential for evaluating sleep quality, yet clinical polysomnography is impractical for continuous home-based monitoring. Ballistocardiography (BCG) enables unobtrusive sleep monitoring through sensors embedded in sleep furniture; however, existing BCG-based approaches either rely on complex physiological feature extraction or employ fixed-parameter signal-to-image transformations that cannot adapt to inter-subject variability. This study proposes a learnable recurrence plot (RP) framework for three-stage sleep classification (Wake, NREM, REM) from single-channel BCG signals. The Learnable RP introduces three innovations: multi-scale phase-space reconstruction at physiologically motivated time delays (τ = 5, 10, 20), differentiable per-scale thresholds optimized end-to-end, and attention-based spatial fusion of multi-scale recurrence maps. The framework was evaluated through 10-fold stratified cross-validation across six backbone architectures using 50 overnight recordings. The Learnable RP consistently outperformed four baseline transformation methods (GAF, MTF, Classical RP, Modified RP), achieving an aggregate mean accuracy of 73.60%, with EfficientNet-B5 reaching 78.91%. and 78.91%. Statistical validation across all 24 pairwise comparisons (4 baselines × 6 backbones) confirmed consistent superiority (all p < 0.001). The proposed framework achieves competitive performance without explicit physiological feature engineering, offering a viable path toward end-to-end unobtrusive sleep monitoring.