Leading researchers in robotics will develop an autonomous robot to be trained using artificial intelligence to find and eradicate weeds in organic fields.
In future, it will be possible to see an autonomous robot the size of a large lawnmower driving around the fields of organic dairy farmers on its own. Here it will look for the weed ‘common dock’, which poses a major problem for farmers, as it takes up space in the field meant for grass for the cows.
Once the robot has found the weed, its next task is to combat it with one of the two instruments it is equipped with. The instrument is partly a laser and partly a device that can give the plant an electric shock. In the long term, the robot will only be equipped with one of the two tools, depending on what turns out to be the most successful in weed control.
DTU’s expertise in autonomous systems
The project includes Wageningen University & Research in the Netherlands, which is the world’s leading university in developing new technological solutions for agriculture. They have invited DTU Electrical Engineering to contribute their expertise in autonomous systems.
“Our part in the project will be to use artificial intelligence to analyse data from the camera and LIDAR sensors installed on the mobile robot. This way, the robot will be able to find and identify the weeds in the field and then eradicate them,” says Associate Professor Lazaros Nalpantidis, who is heading DTU’s work.
The robot makes use of the European Galileo satellites, which can indicate its own position and that of the weeds in the field to within a few centimetres.
“The weed robot is an example of an autonomous system that combines robotics and artificial intelligence, enabling the robot to perform its tasks intelligently without human interference. It can move around the field itself, knows exactly where it is, can find the unwanted weeds and then eradicate them,” explains Lazaros Nalpantidis.
Arla Foods supports the project and will, among other things, help facilitate the communication with organic dairy farmers who can contribute to the development by providing input.
“It’s important to us that our solution is adapted to the wishes and opportunities of its future end users from the outset,” says Lazaros Nalpantidis.
Among other places, the project will also get the opportunity to test its solutions in the Netherlands and France, where other participating partners of the project hail from.
The first autonomous weed robots will be tested in a year’s time, and the entire project is expected to be completed by the end of 2022.