Sensor data from buildings is messy: devices drift, batteries fade, radio links drop packets, and a single bad reading can send an automation rule the wrong way. Before AI can be used to optimise a building, something has to establish which data can be trusted.
AIRE examined AI at both levels: in the horizontal platform that collects and integrates data from many sources, and in the vertical property service that acts on it.
The project showed how AI can be used for quality assurance of sensor data and for extending the automation capabilities of Yggio, and it validated machine learning for predicting energy consumption in buildings.
Just as usefully, it gave the participating companies practical experience of where AI and machine learning are worth applying at all, and where a simpler method does the job. Sensative coordinated the project, funded by Vinnova with 500 000 SEK.
