News

One of the goals of the field robot event is to learn from each other. We are therefore very happy with Kamaro’s initiative to make some of their software public. We expect this to boost our community and encourage collaboration and exchange between teams. It also helps aspiring participants to get started.

Kamaro has published their winning code for crop row navigation and their code for object detection using deep learning.

Which team is next?

CERES II showed a robotic arm in the freestyle task. This arm, completely developed and build by them, is very impressive. It can lift 6 kg and includes a tool head to fit different grippers and actuators. It also measures the force applied by the gripper.

In the first online version of the international Field Robot event team, Robatic was very successful with their robot Bullseye. They won the overall prize. The second prize went to CERES II and Carbonite came in third.

This year’s task 4, requiring robots to remove 10 objects and place them outside the crop on the headland, was the toughest of 18^th Field Robot event. Only 3 teams managed to start in this task, but all failed to perform the required action. CERES II won because it damaged the least plants, Bullseye became second and Carbonite third.

In task 3, the robots scout for weeds and trash in the field. Bullseye detected; eight out of ten. CERES II became second and Carbonite third.

In Task 2, Bullseye did an excellent job. They were the only team to complete the entire field within the limited time of 3 minutes without damaging any plants. Carbonite, the team with no driver’s licence at all, won the second prize. Farmbeast finished in third place.

The first winner of the Basic Navigation task of the online international Field Robot event is Beteigeuze Nova. They managed to finish the entire virtual field in 2:22 minutes, with only four plants damaged! The second place went to Carbonite, the 2019 base navigation winner. Bullseye came in third.

The Field Robot Event is a partner of the DLG Feldtage event. The DLG is streaming this event live on its digital platform.

Information on how to participate:
• Attending this virtual event is free of charge and requires a simple registration on DLG’s digital platform.
• Registration is open from June 7.
• The link to the registration is available here.

• The containers for the field robot event are now available at Github! Further information can be found in the README.md.
• The ground texture has been updated. It does not longer contain images of weeds.
• Changes in the task:
  • All tasks: The starting will always be on the right side of the field. (first turn left)
  • Task 2: A file with the code of the path pattern to be followed is made available to robot at the start.
  • Task 3: Assessment differentiates more clearly between detection during run and classification and accuracy of mapped objects.

Now on Git a new world with waste and weeds is available. Check: https://github.com/FieldRobotEvent/Virtual_Field_Robot_Event

Register your team for FRE 2021 by filling out the registration form on our website (through the Event tab in the menu above).

Check our GitHub! New versions of the virtual environment have been released with improvements for each task.

Have a look at the contest and tasks pages. They have been updated for the online event!

A digital Field Robot Event poses some serious challenges but also brings new opportunities. Due to the digital nature of a virtual event, we can now easily collaborate internationally.

Since today Kamaro engineering, a well-known participant of the Field Robot Event, is contributing to the development of the digital simulation environment. Kamaro engineering already made some beautiful 3D meshes of real maize plants (see picture).

We, as Field Robot Event organisation, are very enthusiastic about this collaboration and we would love to see this field robot community grow. If you have questions, ideas, or want to help with the digital environment, contact us using the discussion thread in the git repository.

In order to prepare, if you have only little experience in GAZEBO/ROS, we recommend to follow the link on the “Virtual environment” page.

This will be also the place where we will publish more details about the event soon!

We still suggest a rough preliminary task description like: Basic (1) and advanced navigation (2), an agricultural application (sensing (3) and manipulation (4)) and the freestyle (5).

And remember, in the freestyle – not compulsory – we suggest to show using a live video your real existing machine at home conducting an agricultural task of your choice.

The online contest will be at the same time as the DLG Feldtage from 8^th to 10^th June 2021. We want to continue our collaboration with the DLG. There will be much likely an online transmission during this time also via the Feldtage digital platform.