Tasks

For this task, the robots are navigating autonomously through a real maize field. This task is all about accuracy, smoothness, and speed of the navigation operation between the rows. Within three minutes the robot navigates between the rows. The aim is to cover as much travelled distance as possible. The field consists of parallel rows spaced 0.75 m apart. Some of the plants in the rows may be missing, but not at the beginning or end of any row. The maize plants are expected to have a height of approximately 0.3 – 0.4 m.

Each robot has to start within 1 minute after the signal for the start. The maximum time is 3 minutes for every field robot beginning with the individual start. The robot must follow a sequence of commands provided in a text file. Each command consists of a number and a direction: For the command sequence 1L, 1R, 2L, 3R, the robot will:

1. Turn left and move to the next adjacent row.

2. Turn right and move to the next adjacent row.

3. Skip one row and enter the second row to the left.

4. Skip two rows and enter the third row to the right

The scoring P includes the travelled distance, the number of damaged plants causing a penalty, and a time-dependent, linearly increasing bonus distance if the field robot finishes before the 3-minute time slot has ended:

This task builds upon Task 1. Robots must navigate autonomously through the maize field, traversing every row without skipping. The primary objective is to detect strawberry bushes (considered weeds), which may be positioned arbitrarily to the left or right of the robot. Detection is valid only within a range of −50 cm to +50 cm relative to each strawberry bush. If the valid detection areas of multiple bushes overlap, they are additive and merged into a union area.

Robots must navigate each row consecutively without skipping. It must detect strawberry bushes positioned to the left or right within the specified valid detection area. Teams must specify to the judges how the robot signals detected strawberry bushes (e.g., as weeds), using methods such as lights, voice, or other clear means.

The scoring system rewards precise detection within the valid area and penalizes incorrect or missed detections. Distance traveled is not considered to avoid giving an advantage to the winners of Task 1.

In this task, the robot operates within a field of dimensions 10 × 10 meters, delimited by hay bales. The field contains five fruit trees, each representing a different fruit: apples, lemons, bananas, grapes, and oranges. These globular artificial trees are positioned at known locations, but with an uncertainty of 20 cm. Thus, the positions are provided but should not be assumed to be exact. The robot starts in the bottom-left corner of the field at the origin (0, 0) and must navigate autonomously while avoiding the trees. The objective is to produce a map identifying the type and location of each tree. During the task, the robot can move freely, using pre-determined strategies or by improvising. For each fruit detected, the robot must signal it to the judges in real time using a signal such as voice, light, or any other paradigm, as part of the evaluation. At the end of the task, the 8 robot must generate a CSV file containing its mapping results in the following format:

fruit_type, x_coordinate, y_coordinate.

The origin (0, 0) corresponds to the bottom-left corner of the field. For each entry, a visual proof of identification must be provided as confirmation to the judges (i.e., an image of the detected fruit). The task duration is limited to 3 minutes, and the robot must complete its mapping within this time.

The scoring P is based on the accuracy of the mapping, comparing the robot’s output to ground truth data. This scoring system rewards precise and complete mapping performance while penalizing missed or incorrect detections.

In this task, the robot operates in the same 10 × 10 meter field as in Task 3. However, the task is conducted under night conditions. The robot is equipped with a UV lamp to identify glowing mushrooms scattered throughout the field. The robot’s objective is to locate as many glowing mushrooms as possible while avoiding collisions that may destroy them. Additionally, there may be non-glowing mushrooms in the field, which must not be mapped. Collision must be avoid for both glowing as non-glowing mushrooms. Mapping non-glowing mushrooms will result in a penalty. At the end of the task, the team must provide a CSV file listing the coordinates of the detected glowing mushrooms in the format: x_coordinate, y_coordinate.

A virtual discrete grid is used to evaluate the task, with soft scoring applied if the provided coordinates are near but not exactly within the correct grid cells. This allows for some flexibility in the robot’s detection accuracy. The task duration is limited to 3 minutes, and the robot must complete its search within this time.

The scoring system encourages teams to balance speed, accuracy, and careful navigation to maximize their score while minimizing penalties. Special attention must be paid to distinguishing between glowing and non-glowing mushrooms to avoid false positives.

In this task, teams are encouraged to present innovative ideas and their implementations on self-chosen topics. Each team must present their idea and explain their solution. Ideas with strong agronomic motivation are particularly welcomed.