The project Internet of Food supports the Sustainable Food Initiative by developing a digital infrastructure to share data and models from different organizations more easily as well as investigating next generation sensing systems to support a more sustainable production process.
The Sustainable Food Initiative aims for sufficient, safe, healthy, affordable and sustainable food for everyone. Directly at the start of this initiative it was already understood that realization of this goal not only requires the development of new product and process technologies, but also the full exploitation of the opportunities of the developing digital world. This relates to maximizing the use of big data, sensing and robotics in the food value chain, for example by the development of data-driven food chains that are sustainable, smart and flexible as well as consumer driven. Or alternatively, by stimulating collaboration between the different partners in the food value chain by the development of a data architecture to translate complex data from different sources into actionable knowledge. The original roadmap is shown below.
Several of the above challenges are tackled by ISPT, NIZO, Symrise and Unilever in collaboration with TUE and WFBR within the Internet of Food project. On the one side, they will develop a digital infrastructure for the (conditional) sharing of data and models from different computational sources and different data and model owners. On the other side, they are investigating the opportunities of next generation sensors and sensing systems in food production. Next to that they are continuously exchanging their experiences in the digital world and thereby contributing to the education of the next generation digital food experts.
In the area of sharing of data and models the Models As A Service (MAAS) architecture has been delivered last summer. This micro-service based architecture permits the sharing of Food models between different companies and organizations. Currently, an upgrade of this architecture is ongoing which will permit the sharing of arbitrary data and models in a much more general way by using a hub and spoke approach with a thin hub (see Figure 2 below). Once set-up users can access data and models from other companies intuitively, removing unwanted inter-organizational barriers such as e.g. reformatting of data. This approach of sharing data using machine readable formats has been realized by the use of ontology's in an early stage of the project.
As mentioned previously the use of advanced sensors will be essential to optimize processing in different steps of the food value chain. In this context a detailed inventory of current and next generation sensors was made. Several of these sensors have been and will be investigated in more detail. Creating clarity on what may be possible using these sensors in future processing lines. For example, one of the cases explored in the project was related to in-line real-time monitoring of droplet size of emulsions during their manufacturing process. To achieve that several optical sensors were explored, for example, in-line microscopy, NIR spectroscopy and in-line multi-spectral technologies. An example of in-line microscopy is shown in Figure 3 below, where microscopy images of an emulsion with different droplet sizes are shown. From this exploration, it can be noted that the technique is highly suitable to capture the changes in the droplet size of emulsions in real-time.
Finally, in the area of educating future digital experts an event is foreseen in November 2021, where the different elements of digitization of the Food value chain will be discussed.
The Internet of Food project will end by December 2021. As the current project was focused on developing the foundations for the Model As A Service approach, it will be important to realize the benefits of data and model sharing in a follow-up project. Discussion on such a follow-up project on a real-life implementation of the MAAS approach and the further development and evaluation of key sensors are ongoing.
If you would like to join the follow-up initiative, please connect with Marjon Navarro (Marjon.email@example.com) for more information