Autonomous Combine Harvests Russian Fields

Russia’s largest agricultural holding company has joined forces with AI development firm Cognitive Technologies to pilot autonomous crop management on 665,000 hectares of land. It is hoped that the technology will help reduce crop losses while also improving farming efficiency.

Rusagro has contracted Cognitive Technologies – the same company that’s testing an autonomous tram in Moscow – to install the Cognitive Agro Pilot self-driving systems into combine harvesters to help manage the holding company’s land reserves in the Belgorod region of Russia’s Central Federal District.

The agreement calls for the system to be capable of safely operating autonomously over plowed or unplowed ground, as well as mowed or unmowed fields, and take appropriate action when objects like other machines, trees, roads, animals and people are detected.

The setup comprises a control system called the Agrodroid, a video camera, a display screen and connecting cables. Human operators will be present in the cabins of the kitted out combines, but the self-driving system will allow them to focus on controlling harvesting systems.

“Usage of the system reduces grain losses while harvesting up to two times,” said Olga Uskova of Cognitive Technologies. “Our highly developed computer vision system allows just one camera to achieve similar results to what most leading international manufacturers get with three or four sensors, including a LiDAR. It also ensures the safe operation of the self-driving harvester in any weather and lighting conditions.”

The autonomous control system can be installed on any harvesters, tractors and sprayers, and the setup comes supplied with a cradle that allows for the Agrodroid brain to be moved from one vehicle to another without having to purchase a new system for each.

Pending successful completion of the autonomous pilot, Rusagro expects the system to be installed on all of its 800 combine harvesters within the next few years.

Autonomous combine set to harvest Russian fields [New Atlas]

 

(Visited 70 times, 1 visits today)