ECG-Based Fatigue Detection in Elderly Individuals Using Machine Learning Models

Fatigue detection in elderly individuals is critical for preventing health complications and enhancing quality of life. This study presents a comprehensive approach to monitoring and classifying fatigue states using electrocardiogram (ECG) signals, leveraging advanced machine learning techniques. The proposed methodology involves five key stages: ECG signal acquisition, preprocessing, feature extraction, dimensionality reduction using Principal Component Analysis (PCA), and classification through recurrent neural network models. Among the evaluated models, the Gated Recurrent Unit (GRU), the Long Short-Term Memory (LSTM) model, and the standard Recurrent Neural Network (RNN) model exhibited the performance, of 98.86%, 97.73%, and 82.76%. respectivily. The results underscore the GRU model’s superior ability to classify fatigue states accurately, highlighting its potential for real-time applications in elderly care. This study emphasizes the importance of robust signal processing and advanced neural architectures in developing efficient fatigue detection systems, paving the way for improved health monitoring solutions for aging populations.