“You see this right here?” Thomas Anken raises a fist with his thumb sticking up as if trying to hitch a ride, “This is the tool farmers use the most in today’s high-technology world.” Anken extends his arm and looks past his thumb to the crowd, which for a moment is transported out of the comfort of the conference hall beside Lake Zurich and into an early morning on a farm in Appenzell. Here the best way to decide which fields to water today depends on the direction of the wind hitting a dirt-stained thumb.
The reasons for the slow uptake of digital technology in farming and food production are multiple: Natural systems are chaotic and have many influencing factors, such as the weather, soils, crop varieties, and pests. And while predictive models can handle extremely complex systems, they require a lot of data. In an industry like agriculture, this data will come from embedded sensors that measure the soil and air. But for now these types of analyses are carried out in the lab and the data is therefore neither real-time nor connected. Another inhibiting factor is farming’s sociocultural milieu, in which many farmers just cannot see how a computer could help them make better decisions than those that century-old traditions have already taught them to make. “To characterize the health status of animals, the farmer’s eye is still the most important sensor,” says Anken.
Despite these impediments, there are several reasons to believe that smart farming techniques will be adopted in the not-so-distant future. First of all, certain concrete advances in agricultural technology are reaching the market. Machines like combines, tractors, and choppers are increasingly equipped with telemetric systems such as GNSS-auto steering. The first robots to give precise dosages of pesticides and herbicides are being developed, and as many as 500 milking robots have been installed in Switzerland.
One promising direction to take could be to develop artificial intelligence systems that employ deep-learning technology—something Google already has its eye on in the agricultural sector. This technology may improve agricultural systems in the near future by creating tools able to handle the complex and multifactor systems found in nature. And as the price of embedded and connected sensors falls, the performance of systems for pest forecasting, fertilization distribution, and weed recognition will certainly be improved significantly when those sensors are combined with self-learning IT systems.
Farmers will continue to rely on their trained eye to quickly tell if their animals or crops are healthy or sick, but perhaps they will soon begin using their thumbs for more than testing wind direction in the morning. They could use them to control robots in the field with a smartphone or to write a text message that says they will be home earlier for dinner because their working day has ended sooner thanks to pervasive smart farming techniques.
Dr. Thomas Anken, Head of Agricultural System Engineering at Agroscope and Dr. Peter Braun, CEO of Swiss Food Research gave a World Café workshop at the 2016 CSEM Business Day in Zurich.Here is a summary of the presentation and group work.