- Are there plans for a few online meetings with all teams before the event?
We have no plans at the moment. If you express an interest, we can of course schedule one
- Is it allowed to have an active connection to the robot to help with computing during the tasks?
We used this approach from 2008 to around 2012, but switched backed to “on the robot” processing due to delay and wifi interferences. Otherwise We see no reason why you could not use this as long as the competition is fare (no manual intervention!!!). So the short answer is yes.
- Could you confirm that the crop rows will not all have the same length and will be terminating with some random offsets? In this case, what is the range of these offsets?
The different crop row length was not one of the challenges we set this year. However, some minimal offset is possible due to different circumstances (around 1 plant).
QUESTIONS FOR TASK 1:
QUESTIONS FOR TASK 2:
- How much water do we need per plant?
Not much. Just to show the principle. I think most robots will run out of water during the task and this was the reason we added indicators as it will probably also be easier for the jury to asses.
- Do we need to spray the leaves or can we water the base of the plants as well?
You can spray the whole plant, just keep in mind to stop the spraying if plants are not present (this is what will count the most).
- Do we need to spray continuously and only stop on turns and gaps or can we spray in an interval to preserve water?
It is advisable to stop the spraying when it is not needed, to preserve the liquid. As it should be quite easy to stop the spraying during turning procedure, we did not think of minus points for this and I do not think they are needed as the robot / team that does this will probably end up with less points as it will ran out of water.
As far as the “full plant rows” goes, I would recommend to keep the spraying on, otherwise the jury might evaluate this with minus points due to the following part of the rules:
– in case the robot fails to detect the plants in the middle of row where the plants are not missing, evident as the robot will stop spraying and / or the light indicator will turn off, this will result in penalty points as in the case of damaged plants (e. g. one damaged plant per one false positive action).”
– we will add a note here: “Some robots might be quite accurate to detect where the plants are and where they are missing. This might result in constant on / off spraying situation where the plants are present. In order to avoid penalty points, the teams are advised to keep spraying / indicator on until the next plant (if present).”
- Can we use a trailer only on this task to be able to spray more water?
Of course. The use of trailer is permitted.
QUESTIONS FOR TASK 3 & 4:
- Do the images have an white border and if yes how big is it?
1. The images will have a white background. How big it is it depends on the motives (most likely on A3 format)
- We are training our detection model, but wondered when we will get the images needed for task. Will that be at the event itself?
2. Check the news in the fieldrobot website. We posted sample images a fee weeks ago. For the 4th task those will be used, while for the 3rd you have to bring 3 with you.
- In task 4, will the obstacle be placed at 1.5 m from the robot, as in task 3? If so, is it ok if the robot stops with a small delay (while still before touching the obstacle)?
3. In task 4 the obstacles will be placed before the start. So, the robots will be moving and 1.5 m depends on where the robot will stop. Hopefully before hitting the obstacle
- .For task 4, the rulebook states “the pictures of the obstacles will be provided to the teams before the event”: when will these pictures be provided?
4. The pictures are the same as for task 3 and were published on FRE webside around a month ago.
- How images will be changed. Will the A3 leave the row before a new image comes in the robots view? Or will you hold multiple images and drop them down each time?
5. You can (probably) detect when an obstacle is in front of the robot, wait a bit (maybe – couple of seconds?) and do the classification (once). The number of images could be up to 3 times the number of competing teams (if this number is high, we will reduce the number of images, so the task will not take so long, by using just two or one image from each team). The idea is to remove the A3 obstacle with each change of the image, so the robot can detect the difference.
- There is only a maximum time of 10 seconds in changing images and no minimum time?
6. We did not limit the minimum time.
QUESTIONS FOR TASK 5: