PlantMap is a project of the Lower Saxony DFKI Laboratory at the Osnabrück site and comprises a start-up project within the EXIST research transfer. The Plan-based Robot Control research group, led by Prof. Dr. Joachim Hertzberg, is concerned with autonomous mobile robots that have only incomplete control and knowledge in their environment.

Applications of basic research in different domains range from the development of long-term autonomous robots to the integration of the corresponding robotics technology to support humans in current semi-automated processes. The most prominent application domain is agriculture: as a data- and knowledge-intensive, highly digitized field with uncertainty and dynamics, it enables AI technology to be put to good use. In the “Agrotech Valley” around Osnabrück, our research transfer is anchored in the existing ecosystem of innovative medium-sized agritech companies and research partners in the agricultural and food industry.

The EXIST Transfer of Research is a funding program of the German Federal Ministry for Economic Affairs and Energy and is co-financed by the European Social Fund (ESF) and supports both necessary development work to prove the technical feasibility of research-based start-up ideas and necessary preparations for the start of the company. The EXIST program is divided into two funding phases and generally extends over a funding period of three years.

The PlantMap project includes a high temporal and spatial resolution three-dimensional plant map of individual plants. Supporting tools and technologies could encourage entry and transition to such growing methods, as well as address the shortage of skilled, well-trained vegetable gardeners. With this future-oriented motivation, the idea for the PlantMap project emerged around a five-member founding team, which is developing the foundations for this vision in the research area of plan-based robot control at DFKI. The navigation and control software Move Base Flex, which is used worldwide and developed by us, as well as the extension for multi-layered three-dimensional impassable environments (mesh navigation) allows autonomous navigation in steep and unstructured terrain, such as vegetable, fruit or market gardens and agroforestry farms. An autonomous robot can thus track the condition and development of individual plants, their shape and phenotype, and important plant parameters on a daily basis, ultimately recommending courses of action. With the help of AI training data and our technology in combination with bio-intensive agriculture, we contribute to a significant optimization of ecological and bio-intensive cultivation methods. Systematically capturing agronomy know-how for small-scale and bio-intensive micro-farming is critical. Knowledge of crop rotations and plant neighborhoods are the foundation for polydiverse farming practices, natural protection against plant diseases, and a healthy ecosystem.

Through the PlantMap, the influence of environmental factors, gene activity, and phenotype can be measured and targeted. These plant parameters as well as training data can be derived from autonomously recorded data. Training data can be used not only to automatically detect plants, their developmental stages and determine plant diseases, but also as a basis to derive recommended actions for vegetable gardeners, plant breeders and farmers.

 

Funding source

The BMWK funding program EXIST Research Transfer is co-funded by the European Social Fund (ESF). The ESF is one of the European Structural and Investment Funds (further information).