EDIoT will develop customised low energy computing versions of SotA energy disaggregation algorithms to run on IoT smart meter devices. Energy management optimisation in CPSs like buildings and smartgrids requires knowledge of the energy consumption daily routines on device level and this information is extracted by Energy Disaggregation (ED) methods. ED is based on machine learning algorithms requiring significant computational power and thus energy samples are transmitted to the cloud for processing rather than running the ED algorithms on the IoT smart meter device. In EDIoT we will develop new/customised light ED algorithms in order to be able to run on the IoT smart meter device.
Through EDIoT MEAZON will enhance their existing smart metering product with on-device ED, which will allow them offer innovative IoT and CPS energy management services to existing and new clients.
UH has developed SotA algorithms for ED, validated on internationally recognised benchmark datasets, and will benefit from EDIoT by getting first-hand experience and knowledge of transforming ED demonstrated technology to a commercial system.