Published Paper Details:

Cognitive fatigue is a significant psychological and physiological condition that impairs an individual's ability to concentrate, process information, and perform tasks effectively. In modern digital environments, prolonged computer usage across professional and academic settings has substantially increased mental workload, leading to rising concerns about undetected fatigue. Traditional detection methods rely on intrusive physiological sensors such as electroencephalography (EEG) devices, eye-tracking systems, and heart rate monitors. These approaches are expensive, uncomfortable, and unsuitable for continuous real-world deployment. This paper proposes an AI-Powered Cognitive Fatigue Detection System using Keystroke Dynamics as a non-intrusive behavioral biometric. Keystroke behavioral features including typing speed, key hold time, inter-key latency, error frequency, and pause duration are analyzed using machine learning classifiers: Logistic Regression, Random Forest, Support Vector Machine, and an Artificial Neural Network. The weighted ensemble achieves a binary classification accuracy of 92.0%, with the ANN achieving 91.3% individually. The system operates with total pipeline latency under 40 milliseconds and CPU utilization below 3%, making it suitable for deployment as a lightweight background monitoring application. A Streamlit-based interactive dashboard provides real-time fatigue scores, session analytics, and evidence-based cognitive reactivation recommendations.