Building block “Stress Observer” allows to evaluate stress level (acute stress) of a person in real-time. It was initially developed for the European project “BonVoyage” to evaluate traveller’s stress depending on used means of transport. Physiological measurements, such as galvanic skin response, blood volume pressure or temperature are collected by an off-the-shelf connected watch worn by the person. These measurements are sent by Bluetooth link and processed using machine learning models to calculate a stress score. The building block can be run and displayed in real-time by an Android application but can also be run on a PC as it is developed in Python language. Actually, the system uses the Empatica watch as wearable, which needs to have an internet connection in the surrounding area to launch measurements. In case of smartphone use to receive data, Wi-Fi connection or PC connection is needed to download measurement files for further analysis. The system’s advantages are real-time, wearability and mobility. Thus stress level of the operator could be estimated during all different working tasks without obstructing in ambulatory scenarios. Detecting the operator’s stress when collaborating with robots allows to adapt robots’ behaviour and then enhance the well-being of the operator. Several commercial smartwatches which give stress indicators already exist. However, they do not provide real-time assessment of stress. Moreover, it is not possible to identify if those smartwatches provide acute or chronic stress. With this building block, acute stress level, which is relevant during robot and human interactions, is provided. As the temporal resolution is higher with this solution, the consequences of the robot’s actions on human stress could be accurately analyzed. This building block can be used in situations where there is a will to improve physical and mental working conditions and the safety of operators. 
For more information, please contact:
BB1: Stress Observer [CEA]
Artificial Intelligence