A main obstacle to deploy body-worn sensor-based wearable systems in free-living settings is the limited battery life. Runtime of devices such miniature accessories and garment can barely cover the time of a daily activity. A continuous long-term behavioural monitoring system requires energy to sample sensor data, to store or transmit them, but also to perform on-line processing, data abstraction and, in some application, interaction with the user.
Behavioural events are sparse in time. Activities are spaced out by “uninteresting” periods. Continuous sampling is a non-optimised solution. The dilemma is how to efficiently trading-off between quality of information, by perceiving meaningful features from the environment, and saving on-board energy.
By actively changing the duty-cycle the sensor is kept in sleep mode most of the time and active state is requested in response to some interesting activity.
"Saving energy on wrist-mounted inertial sensors by motion-adaptive duty-cycling in free-living", Proceedings of the 2018 IEEE 15th International Conference on Wearable and Implantable Body Sensor Networks (BSN '18), IEEE, March 2018.,
This research project is partially funded by the European Union H2020 MSCA ITN ACROSSING project (GA no. 616757).