Our research projects

Photo: Manuel Gutjahr

Project

Title
DeepFarmBots - MIKIM / KI-basierte Objekterkennung und -klassifizierung, Schutzgehäuse und landwirtschaftliche Feldversuche
Acronym
MIKIM - DeepFarmBots
Start
01.04.2023
End
30.09.2025
Coordinating Institute
OndoSense GmbH

Project team ATB
Allocated to research program
Summary
In the MIKIM - DeepFarmBots project, an agricultural MIMO radar system with AI-based signal processing is being developed that achieves improved angular and distance resolution and is combined with AI-based object recognition that can automatically detect and classify metal rods in maize fields as well as hidden objects in grassland and cereal crops. This is intended to address a hitherto unsolved problem that arises during the harvesting of maize, grain or during mowing due to objects hidden in the crop (metal parts, large stones) - these cause severe damage to combine harvesters or mowers and thus lead to high costs and downtimes. The project results will initially flow into a standalone warning system; later, in cooperation with machine manufacturers, integration into the on-board electronics of agricultural machines and field robots can take place. Within the project, the ATB is conducting extensive trials in the field laboratory for digital agriculture at the Marquardt site and at the LVAT at the Groß-Kreutz site.

Funding
Bundesministerium für Wirtschaft und Klimaschutz
Funding agency
VDI/VDE Innovation + Technik GmbH
Grant agreement number
16KN103120
Funding framework
ZIM