Digital Farming

Our Research

Research focus: Digitalization in agriculture

In agriculture, there is a high degree of heterogeneity between the individual actors and the systems used. This makes it difficult to network the systems due to a lack of interoperability. Digital farming solutions (e.g. farm management systems, decision support systems, agricultural machinery) also often meet with low user acceptance due to a lack of transparency.
 
Our research is directed at overcoming the described challenges and focuses on software and systems engineering in the area of digital farming and efficient data management for innovative solutions in the food chain. To this end, the consideration of systems in the entire value chain and a requirements analysis of all different actors is essential.

Our location in Rhineland-Palatinate offers an excellent basis due to the versatile coverage of different areas, such as wine, vegetable and grain cultivation, but also the immediate proximity to research and industry.

Our chair cooperates with the Friends of Digital Farming e.V., in which renowned companies along the value chain, such as BASF, John Deere, Hochwald, Aldi Süd, but also agricultural practitioners, such as the Hofgut Neumühle Teaching and Research Institute, are represented. Furthermore, the chair is active in the national AI lighthouse project Nachhaltige Landwirtschaft mittels KI (Sustainable Agriculture Using AI) and collaborates closely with the Smart Farming program of Fraunhofer IESE in Kaiserslautern.

Current projects

Agile research in Digital Farming with a focus on end-use adoption and data management

Funding provided by: Friends of Digital Farming

Contact: Kira Willems, Marc Favier

Duration: 2 years

 

The project takes place in cooperation with the Friends of Digital Farming and deals with the two research directions end-use acceptance and data management. The project follows an agile research approach, which allows to readjust the research every half year without leaving the general research direction.

Farmers have high expectations for digitalization in agriculture, however this opinion is not reflected in the uptake of digital solutions, which to date is comparatively low. Therefore, the project focuses on improving end-use adoption as the first research direction. 

Similarly, the availability of digital data is increasing, both in terms of the types of data and quantity. The different systems provide more and more access to data, and innovative, data-driven business models can be enabled. However, relatively few data-driven business models currently exist in agricultural practice. Therefore, as the second research direction, the project focuses on supporting data-driven solutions.

 

Improving end-use acceptance

To improve end-use acceptance of digital solutions in agriculture, it is particularly important to broadly capture farmers' perspectives on digital solutions. The first step is to build a community of practice (CoP) with end users of digital products in agriculture and to identify initial acceptance factors through exchange with a large group of farmers. By using crowd-based requirements engineering as a method, the requirements can then be elicited (in a partially automated way). The development of an assessment framework enables a uniform evaluation of Digital Farming solutions. Following the elicitation of requirements and acceptance factors, this empirical knowledge on Digital Farming solutions will be made available in a knowledge database.
In doing so, the following research questions are to be answered in the first research cycles:

  • How can the inhibition threshold of farmers to test and use digital solutions be lowered in order to increase technology acceptance?
  • What are the reasons for the low acceptance?
  • How can the acceptance of use in agriculture be measured and improved?

 

Support the development of data-based solutions

For the implementation of data-based Digital Farming solutions, data is usually not only held centrally in one place, but should be transferred, analyzed or transformed along the value chain. Especially in the implementation of the Farm2Fork strategy of the EU, food supply chains will play an important role and should be modeled and analyzed with an end-to-end perspective. By establishing an Ag-Data Management Lab, the development of data-based solutions and their integration into the digital supply chain will be supported. In such an Ag-Data Management Lab, data-based solutions for agriculture can be tested with regard to their expected benefits, but also the expected technical challenges. The Lab to be established should be equipped with tools for modeling, rapid prototyping, and technologies for capturing and sharing information. Technical hurdles are to be quickly identified and typical solution patterns formulated. These hurdles and solution patterns are to be formalized and thus made available for reuse. After an initial setup phase, medium-term focus topics can be, for example, in the area of data and sustainability, or reliable AI. 
The following research questions are to be answered:

  • How can the development of data-driven solutions in agriculture be improved?
  • How can the expected benefits be assessed at an early stage?
  • How can expected technical challenges be identified at an early stage?
  • What do typical solution patterns for these technical challenges look like?

Reliable and resilient AI methods for sustainable agriculture

Funding provided by: Carl Zeiss Foundation

Project partners:

  • Zentrum für Nutzfahrzeugtechnik (ZNT), Kaiserslautern: Prof. Liggesmeyer, Prof. Dörr, Prof. Görges, Prof. Teutsch
  • Umweltwissenschaften, Landau: Prof. Schulz, Jun.-Prof. Bundschuh, Prof. Schäfer

Contact: Marc Favier

 

Already today, farmers have to cope with demanding and diverse tasks. These include complex decision-making processes that require a good understanding of interrelationships and mechanisms of action, as well as the automation of time-consuming workflows. Artificial Intelligence (AI) promises decisive progress and is expected to facilitate the accomplishment of these tasks. In this context, research and development is addressing many individual issues with the aim of finding efficient ways of taking various environmental parameters into account. Application areas include:

  • Field cultivation with autonomous, resilient field robots
  • Multi-criteria decision making
  • Stable and reliable AI methods in the case of temporary loss of data
  • Learning based on simulated data or data artificially reduced in accuracy

The aim of the project is to bring together the research activities on Digital Farming for Sustainable Agriculture at the Kaiserslautern and Landau sites in a structured way and to systematically develop them into an internationally visible research focus. The research partners at Kaiserslautern and Landau can build on a distinctive and interdisciplinary background. In particular, the area of digitalization is already a unique characteristic that will be specifically expanded in the coming years. All participating researchers from the two locations are to be involved in the project via workshops, thus developing the first joint research activities.

Sustainable Milk

Funding: Funded under the 4th funding call of ELER-Verwaltungsbehörde as part of Europäische Innovationspartnerschaft Landwirtschaftliche Produktivität und Nachhaltigkeit (EIP-Agri)

Project partners: Lehr- u. Versuchsanstalt für Viehhaltung Neumühle, TUK, Wahlerhof, Schmiedhof, Hochwald GmbH, John Deere GmbH & CO. KG, BASF SE AG V

Contact: Felix Möhrle

 

Rhineland-Palatinate dairy farms are under increasing pressure as the cornerstones of the dairy industry are in flux. The food trade is imposing increasingly stringent requirements on feeding (e.g. origin, CO2 profile of feed, protection of biodiversity). There is also increasing pressure from legislators, the market and climate change (e.g. extreme weather events). The question is how farms can adapt to these changes and maintain their competitiveness.

The Better Milk project aims to improve the competitiveness and sustainability of farms. Specifically, two practical farms are being supported in meeting the latest requirements of the food trade. In an exact experiment, only regionally grown feed is used for this purpose and the impact on milk quality and the economic footprint is investigated.

The Digital Farming Chair is leading the work package Data Management in particular. In this context, the development of a cockpit is planned to visualize the entire process chain including products and sustainability parameters.

Sustainable Embedded AI

Funding provided by: Carl Zeiss Foundation

Project partners: Prof. Berns, Prof. Dörr, Prof. Ruskowski, Prof. Schöbel, Prof. Dengel, Prof. Stricker, Prof. Lukowicz, Prof. Wehn, Dr. Plociennik

Contact: Mengisti Berihu

Duration: 2022 - 2027

 

We are witnessing a breakthrough in Artificial Intelligence (AI) in recent years, which is becoming prevalent in many applications. Reasons for this include the rapid progress in computer technologies, the availability of large amounts of data, and the development of new algorithms and methods. As part of the rapidly advancing digitalization, AI methods are also increasingly permeating our everyday lives, where they interact closely with humans and their environment as embedded systems. A central aspect of this is environmental perception, where conclusions about complex processes often have to be drawn from heterogeneous and noisy sensor data.

The goal of the Sustainable Embedded AI project is to improve environment perception in AI systems. To this end, various systems, e.g. for observation, care, or maintenance, are being investigated in different application areas. Reduced data volumes and more effective processing could favor decentralized application in so-called edge computing. In the project, this is being researched in the application fields "Smart Factory" and "Smart Farming". The entire process is being considered: from the procurement of training data to network architecture and the training, to optimized, energy-efficient hardware implementation.

The Digital Farming Chair participates as a partner in the Smart Farming application area with a focus on research on Explainable Artificial Intelligence (XAI). In this context, the correlations of artificial intelligence and concepts of explainability on user acceptance are investigated.

Nachhaltige Landwirtschaft mittels KI (English: Sustainable agriculture using AI) (NaLamKI)

Funding provided byBundesministerium für Wirtschaft und Klimaschutz

Project partners: TUK, NT Neue Technologie AG, Fraunhofer Heinrich-Hertz-Institut (HHI), John Deere, Julius-Kühn-Institut (JKI), OptoPrecision GmbH, Planet, Robot Makers GmbH, Universität Hohenheim, Deutsche Landwirtschafts-Gesellschaft e.V. (DLG), Friends of Digital Farming

Contact: Felix Möhrle

Duration: 2021 - 2023

Website: https://nalamki.de

 

The NaLamKI project aims to use digitalization to help agriculture overcome various challenges such as a lack of interoperability, climate change, the increasingly important management of nutrients and crop protection, and the growing shortage of skilled workers. To this end, the efficient networking of farmers and industry based on GAIA-X is intended to promote their cooperation. For example, data provided by farmers can be transmitted as training data for artificial intelligence to industry partners, who in turn use their expertise to offer farmers decision-making aids and make the data interpretable. Exemplary fields of application are the detection of plant diseases or the prediction of yield development. To ensure acceptance of use, the solutions developed must demonstrably contribute to process optimization and greater sustainability in crop production.

Within the scope of NaLamKI, the Digital Farming Chair is developing a reference architecture that is to be made GAIA-X compliant. For this purpose, a requirements analysis has to determine which data and services are relevant for the platform. On this basis, the components and interfaces between data suppliers, processors, and users are to be specified and prototypically implemented.

Zum Seitenanfang