UZH develops neural sensors for agriculture

Researchers at the University of Zurich(UZH) are developing a neural network to measure humidity, temperature and pH value on agricultural land. The team led by neuroinformatics specialist Chiara De Luca is researching intelligent sensors and chips that not only process the data on site, but also adapt to new situations as learning systems, according to a statement from the university. The advantage of such a local system is that, unlike conventional artificial intelligence (AI) applications, they do not have to deliver large amounts of data to central computers and then reuse the data locally.

Together with research partners from the Politecnico di Torino, De Luca is testing the developed system in an apple orchard. The data collected there is compared with previous data and the accumulated knowledge is translated into irrigation and fertilizer application responses. "Intelligent sensors can be used where data needs to be processed continuously on site and with little energy," Chiara De Luca is quoted as saying in the press release. A further step would be to measure nutrients. However, this would require the development of sensors that could record and analyze complex chemical processes without themselves being susceptible to frequent repairs or replacement.

The project was presented by UZH at the AI-X Summit on 3. and October 4 in the StageOne exhibition hall in Zurich-Oerlikon. ce/ww