Machine Learning for Fatigue Estimation and Prediction “An Introduction Study”

Fatigue is considered as reduced workability and motivation that affects physical, emotional, and mental activeness. It is a critical concern that influences the precision and accurate implementation of some tasks or the emotional condition. Early detection of fatigue onset is crucial, such that preventative or corrective controls may be presented to minimize work-related traumas, the inexact performance of a task that require high-level accuracy, as well as to avoid making a wrong decision as a result of tiredness. Our goal is to create a non-invasive, proactive model for real-time fatigue estimation based on typical features as tremor, heart rate, and blood oxygen saturation. We expect to set up a relation among the handshaking, heart rate, and oxygen level on one side and the weariness onset on the other. We will use a compact high-precision accelerometer to capture the low-frequency physiological tremor and an optical sensor to detect the heart rate and blood oxygen saturation. Intelligent learning algorithm will be used to personalize user characteristics, such as baselines of the tremor, heart rate, and oxygen level.