Final day delivers usable models and practice-based education

The final presentations at VDL demonstrated that this study assignment advances both education and the market. On February 11, Applied Mathematics students from Fontys presented their final project on modelling sensor locations in the greenhouse. Using sensor measurements and AI models, they built a complete temperature profile that helps growers steer their crop and provides companies with a solid basis for development and validation. The technical foundation was strong and gained further depth through questions from the panel, including Joeri van den Hoek of NXTGEN, Jan Enthoven of TomatoWorld and Rick Karreman of 30MHZ. This made the concrete value of the approach clearly visible.

Many greenhouses already collect measurements, but a complete overview is often missing. The students combined multiple sensors with models that estimate the temperature field throughout the space. This resulted in a map showing microclimate differences and visualising developments over time. That step turns data into a foundation for cultivation decisions.

With a more reliable temperature profile, climate and energy can be controlled more precisely. Sensors can be positioned where they contribute most, scenarios can be tested in advance and deviations can be detected earlier. For tech companies, this provides a clear framework for specifications and validation, while growers gain sharper insight into which measurements truly add value. This strengthens mutual understanding and makes implementation a well-founded step.

The panel challenged the students on model selection, calibration and applicability beyond the training range. Which variables matter most, when does an additional sensor help and how do you demonstrate robustness. This interaction required clear argumentation and stronger validation. As a result, usability increased and it became clear where further research adds value.

Students indicated that working in this practical context showed how their mathematics functions in real-world conditions. Companies encountered motivated talent capable of translating technology into cultivation questions. This is the essence of human capital in agrifood, people who understand data, can explain it and apply it responsibly. The result is not a standalone prototype but a shared foundation to build upon.

Here, collaboration takes a workable form. We define challenges together with growers and technology developers, work with shared datasets and clear validation criteria and build step by step from concept to cultivation trial and application. This creates reusable knowledge, lowers thresholds for SMEs and keeps greenhouse practice leading. It is the approach that makes scaling more realistic.