Moscow

Cognitive_Pilot

Cognitive_Technologies

The autonomous control system based on artificial intelligence is designed for installation on agricultural machinery – farm tractors, combine harvesters. The automated system, in addition to the AI system, includes a camera and a calculator with additional interface units with combine harvester’s onboard systems.

The Cognitive Agro Pilot is based on the technology of deep learning for neural networks. The Cognitive Technologies’ team of developers in real-life conditions worked on collecting datasets (the images that contain the main elements of a field environment used for neural network training) and taught neural networks to perceive objects and situations a smart combine harvester will confront during the process of its work and which should be recognized by the artificial intelligence.

AI systems can see different types of boundaries of a field (edge, cutting), eliminating the need in installing extra sensors on the equipment. The solution also allows for adding new types of boundaries (rowing – the shape of planting certain crops, for instance, corn, as well as that of tilled and untilled parts of a field).

Autonomous control system for agricultural machinery

farming

driverless

driverless_cars

artificial_intelligence

Cognitive Agro Pilot

C-pilot comes with a suite of sensors: TOF cameras, radar, lidars, other sensors for detecting objects, and a GLONASS/GPS based positioning sensor.

The data is processed by an onboard minicomputer. The sensors enable 360-degree &nbsp;environment perception. SLAM (simultaneous localization and mapping) technologies convert the geographic coordinates collected by positioning sensors into the car’s local coordinate system. Once the objects detection data and cartographic information is collected, road scene objects and the vehicle become linked to each other.

Static and dynamic objects’ position and movement tracking, near collision calculation, collision warning occurs at the next module.

The calculation of the optimal movement trajectory is made possible by an intelligent system that uses RRT (Rapidly Exploring random three) and an A search algorithm modification. The programs of the next module’s controller can turn the steering wheel at the required angle with the required speed, bring the car to a halt in time and increase the speed of the vehicle.

The system was used to develop the LDWS solution to warn about a car’s veering from a traffic lane, TSR – to identify road signs, FCW – to warn about a collision when a car rides with traffic, PPS – for the pedestrian protection system.

Tech Solutions

Cognitive Agro Pilot

Cognitive Pilot

add your project!
fill out the form

Tech Solutions

Cognitive Agro Pilot

Cognitive Pilot

add your project!fill out the form

add your project!
fill out the form