Increasing food safety using artificial intelligence and sensors
Assessing the quality and freshness of food is a very complex process, and monitoring these attributes poses a huge challenge. Many stages involve multiple factors that could negatively impact product quality. Products are often exposed to contaminants and temperature fluctuations when they are produced, stored and until they are purchased. Throughout this process, the microflora of fresh products is continually changing, with the majority of products undergoing dynamic physicochemical processes. Ongoing monitoring is therefore necessary to ensure that the product’s quality is maintained right up until purchase. However, continual monitoring is limited by current technology and the specialized analyses needed, and requires a great amount of effort on the side of the business.
The objective of the Zukunftslabor2030 project — which is funded by the German Federal Ministry for Food and Agriculture — is to use artificial intelligence (AI) and innovative sensor technology to develop an efficient and sustainable system for monitoring food products so that significant improvements can be made in the areas of consumer protection, consumer information, food quality and safety monitoring, and food waste reduction.
As such, a variety of future scenarios across the food supply chain — from production to consumption — were considered as part of the project. To do so, the project consortium covered the entire research value chain, from laboratory analysis to data science and food law. Fraunhofer IVV is carrying out part of the laboratory research for the consortium, which is made up of eight companies, universities and institutions.
New technologies and processes for monitoring food products
In order to support the interdisciplinary team of data science experts by providing a sufficient amount of reliable data, Fraunhofer IVV is initially conducting investigations on meat samples using gas sensors, GC-IMS, oxygen sensors and hyperspectral cameras. Using cutting-edge gas analysis, these investigations are also looking into small, rapid mobile measurement methods that could potentially be used to guarantee continuous on-site monitoring for food products. Furthermore, project partners from the Max Rubner Institute, the University of Bayreuth and the Bavarian Health and Food Safety Authority are analyzing the samples for molecular and microbiological changes in quality, such as microbial spoilage. The aim is to establish a correlation between microbiology and volatile detectable compounds.
AI-based food monitoring at the core of Zukunftslabor2030
The digital AI-based food monitoring system created in the project uses data from innovative measuring processes such as spectroscopy, mass spectrometry and volatile analysis to learn new information, also updating itself on a continuous basis. By integrating a wide range of measurement data, the most significant chemical, physical and biological processes in food can be digitally illustrated using computer models. The data collected in the cloud forms a digital twin, and each digital twin is just as detailed as the food products themselves. In order to compensate for variations and measurement inaccuracies at setup, a statistical probability statement is issued for the state of individual foods using the digital twin. The results can be used to predict how the quality of products will change according to the actual storage conditions, among other things.
An interdisciplinary team of experts
The interdisciplinary team of experts is combining their comprehensive expertise from the areas of mass spectrometry, optical spectroscopy, next generation sequencing, sensor systems for volatile components and food law with knowledge from the fields of data handling, modelling, data science and AI.
AI-based quality monitoring for a range of applications
AI-based monitoring is a very useful tool for the food industry in particular, thanks to its continuous, real-time quality monitoring, as it uses a digital, dynamic likeness of the product to transfer benefits directly to trade and consumers alike. Any changes in quality can be caught immediately, guaranteeing the quality of food at purchase and even to the time it reaches the consumer’s plate. In the future, AI-based monitoring could be transferred for use in other fields of predictive/preventive maintenance, for example to monitor machine oil and machine oil levels, gearboxes or other machine components. The use of digital sensor technology and artificial intelligence is relevant in all industries where real-time monitoring can lead to more sustainable and efficient results.