"Neuro-fuzzy modeling of human fatigue"
Neuro-fuzzy modeling of human fatigue Yue Jiao and E. S. Lee Dept of Industrial & Manufacturing Systems Engineering Kansas State University, Manhattan, Kansas 66506 Email: Human fatigue is one of the critical factors influencing health, safety and work performance. However, fatigue is not well defined and is influenced by both physical and psychological factors. In fact, even the measurements or what variables to use to express the degree of fatigue cannot be uniformly defined. There exists a large volume of data with physiological variables such as heart beat and blood flow to measure stress or fatigue. However these data are obtained under extreme conditions with short duration, which is more suited for physical competition. In factory working environment or long distance driving, stress is mild with long duration. Under this mild condition, various measurements or independent variables have been used, some examples are duration, leg swelling, frequency of eye movement, the various physiological measurements and the various psychological measurements or scales such as the Borg rating. Because of the vague and not well-defined nature, neural-fuzzy adaptive network appears to be ideal to model human fatigue. In this work, the adaptive network is used to model human fatigue under both the extreme physical competition condition and the mild factory working condition. The adaptive network not only modeling, but, also, improves the model by learning as more data become available.