Dem@Care: A Study on Activity/Inactivity Detection Algorithms by means of Tyndall wireless inertial measurement units
The increase in average lifespan across the world has been accompanied by a significant growth in the occurrence of dementia with high socio-economic costs. The development of Personal Health Systems (PHSs) will enable people with dementia to maintain independence and inclusion in society, thus improving their overall quality of life, which is the final purpose for Dem@Care project. In this regard, wireless inertial measurement units (WIMU) devices make it possible to gather data and monitor motion of people with dementia highlighting the evolution of the disease and, therefore, providing diagnostic information for the realization of the PHSs.
The work described will contribute to a better understanding of the motion/non-motion detection which is being validated on people with dementia in two different hospitals in Nice and Thessaloniki, both of them partners in Dem@Care project.
Dr Paul Galvin : firstname.lastname@example.org
WIMUs (wireless inertial measurement units) are being employed to capture information through advanced algorithms, in collaboration with Dr Eamonn Newmann in Dublin City University, such as:
- motion and non-motion events of the wearer,
- changes in orientation and in posture (e.g., sitting, standing, walking, lying),
- time taken to complete a change of posture,
- gait analysis.
The goal of this work is to show the progress in activity/inactivity events detection. Several algorithms (such as, SHOD, AMVD, AMD, ARED) and their modifications have been taken into account and tested on a data set gathered in a lab environment. Their characteristics have been analyzed and compared in order to highlight eventual drawbacks, as well as to determine the best performing method. The accuracy obtained by one of the mentioned algorithms (AMVD variant) reaches over 96%.
This research was enabled as a strategic research project with connected health through internal Tyndall funding.