Urban Transport

Rusagro employs autonomous combine fleet for bumper Russian harvest

Rusagro employs autonomous combine fleet for bumper Russian harvest
The Cognitive Agro Pilot self-driving system will be rolled out to 242 combine harvesters during the 2020/2021 Russian harvest seasons
The Cognitive Agro Pilot self-driving system will be rolled out to 242 combine harvesters during the 2020/2021 Russian harvest seasons
View 4 Images
The Cognitive Agro Pilot is reported to work in all weather and lighting conditions
1/4
The Cognitive Agro Pilot is reported to work in all weather and lighting conditions
A single camera feeds a neural network in real time
2/4
A single camera feeds a neural network in real time
The mandatory human operator can keep an eye on the Cognitive Agro Pilot system while focusing on other aspects of the harvest
3/4
The mandatory human operator can keep an eye on the Cognitive Agro Pilot system while focusing on other aspects of the harvest
The Cognitive Agro Pilot self-driving system will be rolled out to 242 combine harvesters during the 2020/2021 Russian harvest seasons
4/4
The Cognitive Agro Pilot self-driving system will be rolled out to 242 combine harvesters during the 2020/2021 Russian harvest seasons
View gallery - 4 images

Last year, AI development company Cognitive Technologies installed its self-driving systems into combine harvesters operated by Russia's largest agro holding company. The pilot proved successful, and the technology will now be rolled out to 242 combines over the next year or so.

The joint venture between Cognitive, Sherbank and Rusagro will see harvesters being fitted out with the technology during the 2020-2021 harvesting seasons in Belgorod, Tambov, Kursk, Orel, and Primorsky Krai, with the first installation coming next month.

"The use of an autonomous control system during harvesting will minimize the risks of negative human factors and will optimize the use of combine harvesters," said Rusagro's Roman Shkoller. "It’s noteworthy that the industrial introduction of the system fell in a year when a record harvest of wheat is expected in Russia."

A single camera feeds a neural network in real time
A single camera feeds a neural network in real time

The level 3 autonomous driving system utilizes a live feed from a single video camera, which is analyzed using a convolutional neural network to allow the combine to maneuver itself in plowed/unplowed and mowed/unmowed fields. The Cognitive Agro Pilot setup is reported capable of operating in all weather and lighting conditions, doesn't need GPS and includes object detection capabilities. If the vehicle leaves the pre-assigned route for any reason, the system will warn the mandatory human operator in the cabin so that manual corrections can be made.

For much of the time though, the human operator will be monitoring operations in real time, and controlling other aspects of the harvest – such as the threshing process, grain cleaning and setting the angle of the header.

Cognitive reckons that the system – which can also be installed in other agricultural machinery, such as tractors and sprayers – could lead to a reduction in grain costs by as much as 3-5 percent, while significantly reducing losses during harvesting. The company says that it now has a fully autonomous system on its development roadmap.

Source: Cognitive Technologies

View gallery - 4 images
2 comments
2 comments
MQ
AUTOSTEER by another name.
BlueOak
“the system will warn the mandatory human operator in the cabin so that manual corrections can be made.” So it isn't really autonomous - there’s a human in the cab. Hasn’t “autonomous”, by this definition, already been in use in the US for years, using GPS? Why avoid GPS? It isn’t expensive.