Field sensors fail quietly. A moisture probe pushed out of the soil, a temperature sensor in direct sun, a battery on its last week: all keep reporting numbers that look plausible. Decisions about irrigation, fertiliser or harvest made on those numbers are worse than decisions made on none.
The project developed methods for qualifying sensor data automatically, so that suspect measurements are identified before they reach the advice a farmer acts on.
Sensative led the project with Agricam as partner. Data qualification later became a theme in Sensative’s platform work more broadly, including the AIRE project on AI-based quality assurance of sensor data in Yggio.
The main result was AI-based detection of data anomalies in Yggio: algorithms that flag a sensor once it starts reporting differently from the sensors around it or from its own history. The project also added support for AI algorithms in Yggio’s rules.
The project is funded by the European Innovation Partnership for Agricultural Productivity and Sustainability (EIP-AGRI) with support from Jordbruksverket, under the Swedish rural development programme 2014–2020.
