The objective of this project is to evaluate the fatigue of the video display terminal (VDT) users by monitoring the multi-physiological parameters through the designed and integrated wearable device. The evaluated fatigue condition of the VDT user can prevent the related syndrome by using the guide of fatigue relief. As the first stage of this project, this paper investigates the real time human fatigue detection based on Electroencephalography (EEC), Electrocardiograph (ECG), Electrooculography (EOG) and Saturation of Peripheral oxygen(Sp02) data to test our concept fin- the preliminary study. We selected and extracted different types of features base on the collected multi-physiological signals. We used the unsupervised learning method k-medoids which is the modifications of the k-means clustering to help us to classify the fatigue condition of the video display terminal user. More importantly, we tested our concept by using 25-subject full overnight multi-physiological data, proposed features and classification methods to monitor the fatigue recovery during sleeping. The results validate our concept and show the evaluated human performance (fatigue) condition becomes recovery during sleeping clearly. It proves that the proposed system can monitor the human performance (fatigue) change for the VDT users and it is able to feedback the value to help them relieve the fatigue and to prevent the related syndrome for the next stage.
