Business researchers optimize potato production from A to Z
At the Hanover Fair from 23 to 27 April, these researchers will be demonstrating their procedures at the Saarland research stand (Hall 2, Stand B 46), looking for partners to further develop them for practical use.
Business Informatics Professor Wolfgang Maaß and his team of researchers want to help farmers and food producers make the most of potato production data - Photo: Oliver Dietze
The potato harvest can be quite rough. While in the past farmers and their families and helpers collected the delicately tubers by hand in the field, today harvesters are used. If something goes wrong there, potatoes will rumble hard over the conveyor belts, hitting stones and colliding with each other. Then the more sensitive tubers get bruised. This can hurt them so much that they are unfit for sale. "If the farmer knows that the potatoes are taking too many hits, he can react. He can adjust the speed of the harvester, for example, or take more earth onto the conveyor belt," explains Dr. Sabine Janzen, Research Associate in the Professor Wolfgang Maaß’ team. The deciding factor is that the farmer must first get that kind of information, in time: on the field, when harvesting.
The business experts of the Saar-Uni and the German Center for Artificial Intelligence are now taking care of this. "Smart Services" is their research focus. Clever farming, smart farming, which includes the potato project, is part of it. "We are researching how to draw value-generating conclusions from data for Industry 4.0. Agriculture is still very traditional. Although it is heavily engineered and digitized, it still does not benefit from the data that is generated here," explains Wolfgang Maaß.
Hannah Stein and Mirco Pyrtek (l.) work together in the team of Saarbrücken's business professor Wolfgang Maaß to collect a wealth of data about the potato. - Photo: Oliver Dietze
His research team makes everything about the potato and its path from the field to the factory very clear for everyone involved; from the farmer, suppliers and producers to the raw material investor. "We research what data is needed. From which aspects can we draw conclusions and make forecasts. So that we can extrapolate how big the losses will be if the driver of the harvester continues driving this way or the other. We will pass on this information in real time,” says professor Maaß. "For our forecasts, we use more data, such as financial data. If you know how the world market moves, you can make real-time forecasts of future earnings," he explains by way of example.
For this purpose, the business researchers create a "digital management shell", a kind of logbook that contains as much information as possible about the potato batch. The researchers also determine how many hits the potatoes have taken with a pain-sensitive artificial tuber. "We use our so-called nPotato on the field," says Sabine Janzen. The potato harvester reaps these and the ‘nPotato’ takes the same way route the harvesting machinery as its real cousins. Using sensors, it detects shocks and rotations. If it's too much, it will warn you. "In order to network physical and virtual objects, to classify collisions and to draw conclusions, we combine methods of machine learning, so-called deep learning methods, with information and communication technologies," says Sabine Janzen.
The researchers link so much data together: When was the potato of which variety harvested where, what is its water content, what will it be used for? They integrate price and financial forecasts and in the future will also use historical data, last year's food logistics processes, weather forecasts as well as expert knowledge of the farmer. In this way, they create a service platform for all who are dealing with tubers. It can be a decision-making aid for the farmer, when to bring the potatoes on the market, whether a batch is for a star chef or rather for cornstarch. Even machines can be adapted, for example to peel a thicker potato peel. Commodity investors can secure purchases through quality seals. "If we look at the data of several farmers together, we can make even wider predictions. This would allow chip makers to choose a different type of potato if the quality of a batch is likely to be such that it will have problems in three months," explains Professor Maass.
For more information:
Universität des Saarlandes
Campus A2 3
66123 Saarbrücken
Deutschland
www.uni-saarland.de
Prof. Dr.-Ing. Wolfgang Maaß
Tel.: 0681 302 64736
E-Mail: wolfgang.maass@iss.uni-saarland.de