Mental Stress Detection Using Multimodal Sensing in a Wireless Body Area Network

This thesis presents the design and implementation of a software framework that allows to collect and store physiological data on small sensor devices attached to a human being (forming a Wireless Body Area Network). It describes how sensor data from two biosensors (GSR and ECG) is collected in a number of field experiments, where subjects are exposed to different emotional situations. The experiments are done together with a group of researchers from the department of psychology at Freie Universitat Berlin, FUB. The collected data is then used to (post factum) extract a set of features, which are used as input for existing machine learning algorithms to study how mental stress of human beings can be detected automatically. The evaluation of the algorithms shows that multimodal sensing can improve the stress detection accuracy.