Trading platform 
Monitoring 
Express applications 
Fork Work 
Service 
News 
Directory 
The company Cognitive Pilot, in which Sber and Cognitive Technologies collaborate, has developed an innovative technology that can improve the accuracy of artificial vision for autopilots using artificial intelligence for agricultural machinery.
"We have applied a neural network mechanism to more accurately define boundaries. By analyzing images from cameras and features of field areas, we can identify the most probable zones containing these boundaries, refine the data about the boundary's presence, and confirm its existence,"
said Gennady Savitsky, the lead developer of the company, to the press service. "In other words, our neural network pays attention even to minor changes in the texture of the field, which helps to more accurately define the boundaries. This is our unique approach that no one has used before."
He also explained that with this technology, the autopilot's neural network can see more than the human eye and determine the necessary boundaries with centimeter precision. Such solutions are especially in demand in tractor tasks, where it is important to define the boundaries between seeded and unseeded areas of the field during planting or between treated and untreated parts with chemicals.
"The importance of developments based on artificial intelligence for navigating agricultural machinery in areas with poor GPS signals or its absence is hard to overestimate. For the agrarian regions of southern Russia, these innovations are a real lifesaver. Such areas are becoming increasingly relevant worldwide.
Factors influencing GPS signal issues can include natural phenomena such as solar flares. On the other hand, AI-based autopilots represent a new generation of technologies.
Thanks to modern sensors and intelligent software, they are able to recognize landmarks for navigation during autopiloting better than humans with maximum precision," noted Natalia Chernysheva, the director of the AgroTech cluster at the Skolkovo Foundation's biomedical technology fund.
