Ilham A.E. Zaeni, Andriana Kusuma Dewi, Putra Wisnu Agung Sucipto, Muhammad Khusairi Osman
Fatigue represents a significant physiological and psychological condition that can adversely affect performance and safety, especially in high-demand settings. This research presents a method for identifying mental fatigue through the analysis of chest ECG signals in the frequency domain. We utilized the FatigueSet dataset, comprising ECG recordings from subjects experiencing cognitive load, to extract frequency features through Fast Fourier Transform (FFT) and subsequently trained a pruned decision tree classifier with these features. The ECG signals underwent preprocessing through bandpass filtering, segmentation, and conversion into a fixedlength spectral representation. A decision tree model was developed to categorize fatigue into three distinct levels: high, medium, and low. The model attained a classification accuracy of 66.15%, demonstrating robust performance in identifying high and low fatigue levels; however, medium fatigue proved more challenging to differentiate. The results indicate that frequency-domain ECG features are suitable for interpretable, real-time fatigue monitoring, with possible applications in wearable health systems and cognitive workload evaluation. © 2025 IEEE.
State University of Malang, Dept. of Electrical Engineering and Informatics, Malang, Indonesia; Universiti Teknologi Mara, Center for Electrical Eng. Studies, Pulau Pinang, Malaysia
