Robots in the greenhouse

In researching the complex trait drought stress, KWS is using a sophisticated system for phenotyping corn and sugarbeet plants.

Sometimes research seems like a video game: The team Ben Gruber guides through the greenhouse using his laptop is ready and waiting at the starting gate. Previously, the KWS biologist had precisely defined the moves and paths of his interactive team: Which robot will leave its position when to then approach one of the numbered plant tables? And in what order will the plants be taken to the various stations in the test series?

Employing forward-looking tactics is essential to keep operations running smoothly. Many plant tables are needed for a collaborative research project involving biologists and physicists at KWS who aim to phenotype corn and sugarbeet plants in a greenhouse. Phenotyping is a process of measuring and analyzing a plant’s external appearance in detail — that is, its phenotype. In this way, both the influence of environmental factors and the influence of the plant’s genetic resources can be recorded.

Digital phenotyping: Robots and AI help in the greenhouse

Gruber and his colleagues are engaged in a topic particularly pressing for farmers: They are researching traits related to drought tolerance. “We still don’t know which genes plants actively switch on when they experience drought stress,” explains the biologist. An in-depth understanding of this is, however, important to effectively support breeding of drought-tolerant varieties in the future, Gruber points out.

Digital solutions have long been in widespread use outside in the field, but now they’re increasingly moving into the protected area of the greenhouse: The flat, at first glance inconspicuous, lowpad robots are reminiscent of mobile lifting platforms. “These devices have proved themselves in online mail-order companies’ logistics centers, and they also work for us here,” Gruber explains. The robots he selected for the test project offer more flexibility than, for example, conveyor belts. This makes them ideal for moving the plant tables, which weigh more than 50 kilograms each, to the various stations in the greenhouse.

Robots help with phenotyping in the greenhouse

The corn and sugarbeet plants are weighed regularly to determine their water consumption over the course of an entire day. The information about transpiration is then combined with other data. A sensor system on the ceiling measures the existing differences in temperature and humidity. “Parameters like these are crucial for the leaf stomata and thus for the plant’s water consumption,” Gruber explains. “With our approach, we hope to better understand how plants react to certain conditions and why, for example, the leaf stomata open at varying rates.”

To clarify this and other issues, such as yield production and yield physiology, KWS researchers carefully examine the plants’ visual traits. The scientists observe, among other things, how the plants’ shape, size and weight change as they develop. So they can do this, the robots transport the plant tables to another station: the PhenoFactory.”

In front of the blue interior wall, the green of the plants stands out particularly clearly. The cameras in the PhenoFactory are automatically triggered once the robots have moved the plant tables into the correct position. The photographs make it possible to see the water content of the plants. The plants are scanned from all sides to obtain information from various perspectives. Gruber says: “The technology enables us to collect data 24 hours a day, 365 days a year. This high test frequency is important for achieving good validity and comparability of the results.”

This comparability was already influenced before the test began — at the point the plants were sown. Selecting the right precultures is therefore important to be able to later study traits like size by looking at plants that have as homogeneous as possible genetically identical traits.

The robots take the plant tables from the PhenoFactory back to their places. Then they head back to their starting area where they wait for their next assignment. “Technology supports us in our research, but it can’t replace know-how and experience,” says Gruber, adding that human expertise would remain indispensable, particularly in analyzing collected data.

Back to the "World of Farming" main page

Your contact

Stephan Krings
Stephan Krings
Send email
CONTACT