Novel Feature based Automatic Identification of Physical Exercise and Motion Noise using Electrodermal Activity

Electrodermal Activity (EDA) has gained interest in the research of physiological signal processing and is a well- known biomedical signal that is being used extensively in the field of physical and psychological stress detection. It has been shown that the analysis of psychological as well as physical stress can be detected using the analysis of EDA. However, the signal is very subjective and varies considerably from person to person. Again, being prone to motion noise, this signal is difficult to analyze in presence of motion. For effective analysis of physiology, it is of much importance to differentiate between physiological signal and motion noise. Information can get lost in presence of motion, and also motion might be misunderstood as physiological signal, which can lead to a poor decision, if not a dangerous consequence. In this study, we have explored an objective parameter to detect steady position EDA from motion noise or physical exercise.