Skip to main content Skip to main navigation

Project | resKIL

Duration:
Resource-efficient AI for Embedded Systems in Agricultural Machines

Resource-efficient AI for Embedded Systems in Agricultural Machines

The application of AI methods appears to be useful in agriculture because the environment is highly dynamic, not fully known and not fully controllable. If AI-based applications are carried out on the mobile work machine (on the edge) on specialised and high-performance hardware systems, there are consequences. The use of high-performance computers is associated with high recurring costs and significant effort in adapting the vehicle architecture. A wireless, reliable and broadband cloud connection is not available in rural areas. These restrictions can hinder the adoption of a technology that has proven to be useful for farmers in various cases. A distributed approach seems to make sense, which is minimally invasive on the machine side and uses a specialised environment on the cloud side. In the resKIL project a software architecture and an AI toolchain will be developed to meet the requirements of the mobile working environment. The prototypes will be evaluated in agricultural practice. The DFKI participates in the tools for efficient data processing/annotation, as well as the scalable, adaptive AI architecture that will work on the previously created data. Possible interfaces to and cooperation with Gaia-X and its agricultural usecase Agri-Gaia are considered

Partners

CLAAS E-Systems GmbH (Koordinator]) CLAAS Selbstfahrende Erntemaschinen GmbH Zauberzeug GmbH Universität Osnabrück - Institut für Informatik - Arbeitsgruppe Eingebettete Softwaresysteme TU Dortmund - Fakultät Statistik - Arbeitsgruppe Mathematische Statistik und industrielle Anwendungen

Publications about the project

Sponsors

BMEL - Federal Ministry of Food and Agriculture

BMEL - Federal Ministry of Food and Agriculture