Cognitive Agro Pilot

Mechanized soil cultivation or crop harvesting is an important technological process that includes a large amount of tasks.

On the one hand, the agro-biological state of plants is quite variable even within the same field. There is a big variability of haulm stand density (19.2%), the height of the ear (8.3%), the yield of straw (30%) and other parameters even within one field. In order to avoid losses, it is necessary to constantly adjust the parameters of the header and beater drum, control the harvesting speed.

On the other hand, the final phase of maturity is very short – only 3-5 days, then the crop begins to crumble, causing an increase in losses. Therefore, the harvesting period is very short (no more than 15 days), while it is necessary to quickly carry out the machinery with accurate trajectory tracking, taking into account the geometry of the crops and dimensions of processing or harvesting units.

Multiple tasks require constant switching of the mechanizer’s attention, which leads to fatigue growth and a decrease in the efficiency of work, especially under limited time of the harvesting.

One of the solutions to this problem is the use of parallel driving systems based on the geodetic systems, but this approach is not reliable when it comes to releasing the operator from the control of a farm machinery unit. Field geometry often has a complex structure, contains natural obstacles – islets of plants, power poles, large boulders. In addition, there may be other equipment and people on the field. All this requires maneuvering, constant monitoring of the situation in the direction of movement. Meanwhile, a navigation-based driving system requires a refinement correction service, which is not always available in all areas.

Cognitive Agro Pilot system is a robot that provides autonomous control of farm machinery with vehicle speed up to 15 kmph (tractor, sprayer, harvester). The basic technology of the system is the analysis of images coming from on-board video cameras via computer vision algorithms. The system is able to “understand” the types and positions of objects along the way, build trajectories and give control commands for the necessary maneuvers to be made.

The main functions of the system are:

  • Control of the vehicle’s movement along the edge (sloped crop, cultivated land);
  • Movement control by row (corn, sunflower);
  • Movement control by windrows;
  • U – turn after passing of harvested crop edge or cultivated area. Including possible skipping of the header width or rolls;
  • Stationary obstacles detour;
  • Emergency breaking in case of collision with other equipment or people.

The system hardware is composed of the following components:

  • On-board front view video camera;
  • Onboard computing unit.

The system software solves the following complex of tasks:

  • Detection of object types during operation. Windrows, rows, edges, obstacles, other machinery and people, poles, treated and untreated part of the field, the trail for the sprayer, the edge of the field are recognized and allocated as individual objects.
  • Positioning of detected objects relative to a unit of equipment under the control of the system – a detailed obstacle map is built up around the robot.
  • Building of motion trajectory is performed taking into account the robot dimension and installed units, according to the criterion of maximum safety for people and other machinery.
  • Transmission of commands to the controls of the equipment – the signals for controlling the rotation and speed are transmitted to the corresponding technology units.

It is important to note the high accuracy of following the course taking into account the geometry of the edge, windrows or row. In this case, the system operator can set the desired indentation. The accuracy is 10 cm.

Agricultural technological cycle characteristics are such that harvesters are used during a very limited cropping period. In order to increase the efficiency of the system, we made it possible to quickly re-equip the C-Pilot robot on different agricultural machinery.

During the harvesting and sowing, the tractor is running under the control of the system, during the cropping period it quickly moves to the combine.

We are confident that our intelligent Cognitive Agro Pilot system will become a reliable assistant in your agri-business.

Object detection
Hayroll detection