Towards sustainable food quality with KINLI

Project timeline: Sep 2022 – Aug 2025

Funding: German Federal Ministry of Food and Agriculture (1,6 M€)

Project coordinator: Fraunhofer FIT

Tasks of FIT: requirement analysis, data hub evaluation, visibility and publication of research findings

Current supply chains aim to enable transparent tracing of products and fast reactions. Meanwhile, at the core of KINLI’s aspirations lies proactive quality control. The project stands for ‘Artificial Intelligence for sustainable food quality in supply chains‘ (‘Künstliche Intelligenz für nachhaltige Lebensmittelqualität in Lieferketten‘ in German) and works towards a sitewide data platform. Competent partners in economy and science contribute to this endeavor.


Rising challenges for food production

The course of climate change requires manufacturers to take responsibility for sustainable foodstuffs. As part of this process, appropriate husbandry conditions gain importance: However, only complex measures manage to secure quality and animal welfare conditions so far. KINLI would like to simplify food control across supply chains and save valuable resources. Through an early recognition of possible shortcomings, the project ensures top-quality products and supports sustainability guidelines by Germany’s Federal Ministry of Food and Agriculture.

Predictive approach thanks to Artificial Intelligence

A collection of numerous text files and visual data called data lake serves as the foundation of the central data hub. Economic practice partners as well as the already established cloud platform CERES provide it with measurement data. From there, selected points of intersection apply AI services to help diagnose problems before they occur. Deep learning techniques cluster regularities within training data. Before implementation, this system is tested for weak points to facilitate real-time data processing. The tasks of Fraunhofer FIT involve overall project coordination, the development of KINLI’s platform architecture and an overarching requirement analysis.

Furthermore, the project’s ability to function in real-life situations is supported by strong cooperations. Sauels GmbH makes its products available for quality analyses and provides insights into its production processes. At the same time, Kolsert KG supplies central data on the well-being of their animals. The Institute for Machine Learning and Analytics (IMLA) at the Offenburg University of Applied Sciences focuses on evaluating near-real-time data along the supply chain and on developing AI services. The Faculty of Food and Nutrition Sciences at the Niederrhein University of Applied Sciences is responsible for creating a concept for AI in the meat industry. It also examines KINLI’s use cases with the companies involved.

Flexible applications

To ensure technical advantages for the food sector and the meat industry in particular, KINLI includes two practical use cases. Its first area of application covers the guarantee of food quality using examples of cooked ham products by Sauels GmbH. Artificial Intelligence should automatically mark irregularities and make visible the factors which contribute most to exceeding the threshold values. This way, fewer goods that do not meet consumer demands would have to be sorted out. 

Secondly, animal welfare resembles the aim of the other use case. An image data analysis shall judge the health situation of livestock at Kolsert KG. Room gas analyses, questionnaires and other health measures should also be part of the assessment. From this, the project hopes to reduce the use of antibiotics due to illness. 

Efficient connection of supply chains

At the production level, KINLI’s assessment of the risk of defects can help avoid a waste of resources. Thanks to knowledge about individual transport chains, packaging material can also be optimized before products enter the distribution chain. Hence, an overall improvement of the production process in terms of sustainability is possible.