TPU Graduate Develops Neural Network to Manage Greenhouse Microclimates
A Tomsk Polytechnic University graduate has received a 1 million-ruble (about $13,000) grant from the Foundation for Assistance to Innovations to develop the project.
Tomsk Polytechnic University graduate Alexander Sokolov is developing a neural network designed to manage greenhouse microclimates. The system will monitor humidity, air temperature, and atmospheric pressure, while a compact weather station equipped with sensors will be installed outside the greenhouse. Using the collected data, the algorithm will automatically adjust the greenhouse's environmental conditions.
Sokolov received a grant of 1 million rubles from the Foundation for Assistance to Innovations to support the project.
“There are comparable products on the market, but no one has attempted to integrate neural networks into greenhouse climate control,” Sokolov said. “Many developers have built ‘smart’ greenhouses, but those systems rely on linear control logic. For example, if the temperature rises above 20 degrees Celsius, a fan simply switches on or off. We are taking a different approach by training a neural network on a wide range of operating scenarios so that it can manage crop-growing conditions with much greater precision while reducing the workload for farmers.”
