SPHERE researches, designs and manufactures its own, bespoke, wearables. They are designed for long-battery life (since many participants may not be able to commit to frequent recharging of devices), and they are designed to continuously stream data via a network of SPHERE Gateways in the participant's home, rather than storing it locally.
They primarily measure activity using a 3-axis accelerometer, and the SPHERE Gateways also measure signal strength from the wearable, allowing it to be located within the home using a variety of novel indoor localisation techniques, many of which are machine-learning based. This location information gives metrics of mobility (e.g. how many times does a participant move between rooms during the day) but also important contextual information (e.g. knowing someone is in the kitchen is a good initial indication that they might be cooking).
SPHERE has done a lot of work on fusing wearable and silhouette data which can greatly improve estimation of a person's quality of movement. This video, created for SPHERE participants, explains more:
There have been three versions of the SPHERE Wearable:
The first iteration used a coin-cell battery to power the device. This was useful during initial development and testing, but the practicalities of changing batteries every week prevented it being used in deployments.
The second iteration swapped the coin-cell for a rechargeable Lithium Polymer (LiPo) battery that could be topped up using a Qi wireless charging device. These wearables were used throughout the SPHERE 100 Homes project and proved robust in a domestic setting.
The third iteration aimed to make the wearable smaller and more comfortable. After experimenting with a number of case and PCB designs, a shower-proof screw-tight enclosure is now being rolled out to the SPHERE Next Steps projects.
Prof Rob Piechocki, Dr George Oikonomou, Prof Ian Craddock
Find out more about our research via the links below:
McConville, R.; Archer, G.; Craddock, I.; Kozłowski, M.; Piechocki, R.; Pope, J.; Santos-Rodriguez, R.
Future Generation Computer Systems, 2021.
Holmes, M.; Perello Nieto, M.; Song, H.; Tonkin, E. L.; Grant, S.; Flach, P.
Vafeas, A.; Biswas, I.; Fafoutis, X.; Elsts, A.; Craddock, I.; Piechocki, R.; Oikonomou, G.
In Proceedings of the 2020 International Conference on Embedded Wireless Systems and Networks
Twomey, N.; Diethe, T.; Fafoutis, X.; Elsts, A.; McConville, R.; Flach, P.; Craddock, I.
McConville, R.; Archer, G.; Craddock, I.; Horst, H.; Piechocki, R.; Pope, J.; Santos-Rodriguez, R.
In KDD Workshop on Machine Learning for Medicine and Healthcare, 2018