Moscow-based Cognitive Pilot, a developer of AI systems for autonomous vehicles, and Sberbank-Telecom, a supplier of comprehensive telematics solutions and a vehicle telematics platform, have announced a partnership to create a new agtech solution for the Russian and international market. Called Cognitive Sber Agro Analyst (CSAA), the integrated platform will aggregate high-quality smart analytical data and monitor agricultural operations, such as the yields and state of crops, farm equipment modes, fuel consumption rates, and other statistics.
According to a Global Market Insights Inc report, the precision farming market will triple in value from today’s US$4 billion to US$12 billion by 2025. Precision farming in Russia is in its infancy and is expected to boom in the next three to four years to be worth some US$210-280 million, analysts say.
The companies are scheduled to roll out an integrated SaaS solution based on Sberbank-Telecom’s Smart Telematics Platform for comprehensive monitoring of equipment and machinery, combined with the Cognitive Agro Pilot software and hardware suite, based on Cognitive Pilot’s AI that allows operators to run agricultural machinery automatically.
“Jointly with Sberbank-Telecom we’re set to market a unique solution in and outside Russia,” said Olga Uskova, CEO, Cognitive Pilot (a joint venture between Sberbank and Cognitive Technologies). “Based on a telematics cloud platform and AI technologies, it is designed to improve the efficiency of agricultural enterprises.”
Ruslan Gurdzhiyan, CEO, Sberbank-Telecom, added: “Collaborating with Cognitive Pilot, we are focused on expanding the capabilities of our platform and creating a comprehensive SaaS service for customers to monitor the quality of agricultural machinery’s work, increase harvesting productivity, cut losses, and create relevant field yield maps based on the data obtained directly from the equipment.”
The platform features Sberbank-Telecom’s telematics cloud platform, and the Cognitive Agro Pilot suite, which includes a set of sensors (video cameras and radars), a display, an onboard computing unit, and connecting cables.
While auto-piloting the farm equipment, the Cognitive Agro Pilot system collects comprehensive data from sensors on the status of agricultural processes and sends them to the Sberbank-Telecom cloud platform, where user-friendly control and monitoring tools generate digital field maps and analytical reports, as required. Users will get an effective tool for controlling agricultural processes and monitoring the work of the equipment fleet, which is vital for building yield maps and detecting potential issues (for example, how much a harvesting header has captured and harvested, the quality of the grain in a bunker, grain harvest losses, how much grain trucks failed to deliver to the elevator, etc).
The system also delivers analytical data about the parameters of a particular agricultural process and whether its comprehensive control is possible. The use of computer vision technologies for monitoring purposes will also help identify emergencies and incidents, such as the dangerous proximity of a piece of equipment with an obstacle.
Cognitive Pilot and Sberbank-Telecom believe the joint solution will save up to 30% of costs an agricultural enterprise incurs by cutting unrelated expenses, analysing bottlenecks, securing comprehensive control, and boosting the efficiency and transparency of business processes in agriculture.
The partners also believe the use of computer vision technology, in contrast to the use of GPS navigation, will help to collect data of a different quality. This includes movement safety control with the possibility of generating analytical and monitoring reports about the safety of agricultural machinery and maintaining operating modes that would be optimal from this point of view.
Computer vision systems also allow stakeholders to interpret a particular situation correctly and send accurate data about a particular process to the cloud platform. For example, when a harvester header fails to capture a significant cropping area, a ‘blind’ system usually interprets this as a routine situation, only with less input data corresponding to the area missed by the harvester header. A computer vision-based solution, like a person, controls and detects errors while harvesting, which will be taken into account and documented when generating an analytical report.
The parties are currently finishing their work on the joint pilot solution, with field tests scheduled to begin in Moscow later this month.
