Geoscan has for the first time demonstrated operation of a drone with the neural network on board at the engineering competition of the Archipelago-2025 project and educational workshop, the press service of the Russian company told TASS.

"Geoscan company has for the first time demonstrated neural network operation within the framework of the event in real flight conditions on board of the Pioneer Mini 2, the first serial Russian educational quadcopter with artificial intelligence on board," Geoscan said.

The unmanned aerial vehicle completed tasks of objects’ recognition, following a predetermined route and launch of own scenarios without a server connection, the company noted. "Data processing right on the board provided for the high rate of response and operational accuracy in dynamically changing conditions, which is important when training in operation of real unmanned aerial vehicle systems," it stressed.

The company sees dramatic growth of interest in engineering tasks with AI use, especially among schoolchildren, Deputy CEO of Geoscan Moscow Daniil Zolotnik said. "This is exactly why we create open and affordable instruments for practical studies. The Pioneer platform provides an opportunity to go along the path from first strings of the code to solutions for real flights," he noted.