Assignments to perform during the Field Robot Event

The international Field Robot Event is an annual contest on a, in 2021 virtual, agricultural field, where students and their supervisors compete within several tasks in autonomous navigation and other operations. The FRE has been founded by the Wageningen University in 2003 in order to motivate students to develop autonomous field robots. Now we are looking forward to the 18th event and hope to enjoy creative and functional solutions. The agricultural tasks will be challenging for the robots and their students, but behind engineering skills the organisation wants to promote meeting international colleagues ‐ and of course having fun during the contest!

This year the competition will be conducted virtually in a simulation of ROS Gazebo. It is an open-source 3D robotics simulator. Gazebo supports codes for sensor simulation and actuator control. It provides realistic rendering of environments including high-quality lighting, shadows, and textures. It can model sensors that “see” the simulated environment, such as laser range finders, cameras (including wide-angle), Kinect style sensors, etc.

Task 1 – Basic navigation

For this task, the robots are navigating autonomously through a maize field. Within three minutes, the robot has to navigate through curved (!) rows. The aim is to cover as much distance as possible. On the headland, the robot has to turn and return in the adjacent row. This task is all about accuracy, smoothness and speed of the navigation operation between the rows.The starting is on the left side of the field (first turn is right) or on the right side (first turn is left). This is not a choice of the team but of the officials. Therefore, the robots should be able to perform for both options. A headland width of 2 meters free of obstacles (bare soil) will be available for turning operations. The headland will be indicated by a fence or ditch or similar (3D object).

Random stones are placed along the path to represent a realistic field scenario. The stones are not exceeding 25 mm from the average ground level. The stones may be small pebbles (diameter <25 mm) laid in the ground and large rocks that protrude (maximally 25 mm) from the ground. In other words, typical outdoor abilities as defined by machine ground clearance and to climb over small obstacles are required.There will be no gaps in row entries as well as at the end of the rows. The ends of the rows may not be in the same line.

Task 2 – Advanced navigation

For this task, the robots are navigating autonomously. Under real field conditions, crop plant growth is not uniform. Furthermore, sometimes the crop rows are not even parallel. We will mimic these field conditions in the second task. The general rules here are the same as in task 1. No large obstacles in the field, but more challenging terrain in comparison to task 1. The robots shall achieve as much distance as possible within the duration of the task while navigating between straight (!) rows of maize plants, but the robots have to follow a certain predefined path pattern across the field (picture 2 at the end of this text).

The robot must drive the paths in a given order provided by the organisers. The code of the path pattern through the maize field is done as follows: S means START, L means LEFT hand turn, R means RIGHT hand turn and F means FINISH. The number before the L or R represents the row that has to be entered after the turn. Therefore, 2L means: Enter the second row after a left-hand turn, 3R means: Enter the third row after a right hand turn. The code for a path pattern, for example, may be given as: S – 3L – 2L – 2R – 1R – 5L – F. The code of the path pattern to be followed is made available to the teams two hours before the start and one hour before “submission of code”. This means one hour will be left for the teams to check the required navigation pattern in the test environments.

Random stones are placed in the same way as within task 1. Additionally, at some locations, plants will be missing (row gaps) at either one or both sides with a maximum length of 1 meter. There will be no gaps in row entries as well as at the end of the rows. The ends of the rows may not be in the same line.

Task 3 – Field mapping

The robots shall detect objects as weeds and beer cans (example for waste) and map or geo-reference them. The coordinate system shall be locally in horizontal field dimensions. The reference point will be a point marked by a QR code for camera detection and a pole for e.g. LIDAR detection. Task 3 is conducted in an environment similar to task 2. Nevertheless, good row navigation is required. There will be nine (9) objects in total distributed across the virtual field.

The robot has to generate a file (*.txt) with detected objects and their coordinates relative to the given reference point. Each object should be reported on one line in the file together with the coordinates x and y in horizontal plane in meters with 3 decimal points. Extra points can be obtained for object classification (weed or waste).

Objects are realistic weeds and cans e.g. of beer with different brands and colours. The objects will be placed randomly across the field. So they can be in between and in the rows. No objects are located on the headlands.

task 4 – removal of objects

For this task, the robots are navigating autonomously. The robots shall remove 10 objects and place them outside the crop area on the headland. Task 4 is conducted on the area used in task 2 and task 3 with straight (!) rows. The exact relative position coordinates of the objects will be provided beforehand as a realistic scenario using a drone with a positioning system and a camera. The object type will also be provided. The waste / cans (5 objects) should be placed on one headland and the weeds (5 objects) on the other headland. There will be an indication of a field reference point using a QR code and a pole.

The weeds and beer cans will be randomly distributed across the field. So they can be in between and in the rows. No objects are located on the headlands. The weeds and cans will not be connected to the soil, but will differ in size and mass.

Task 5 – freestyle

Teams are invited to let their real robot perform a freestyle operation at their home institution. The explanation as well as the performance shall be transmitted online via internet to the jury and the spectators. One team member has to explain the idea and the machine. Comments during the robot’s performance are also welcome.

Creativity and fun are required for this task as well as an application-oriented performance. The freestyle task should be related to an agricultural application. Teams will have a time limit of five minutes for the presentation including the robot’s performance.