Suggestion for future rules

RoboCupJunior Soccer
#1

As far as I know last year there was an idea to post interesting rule suggestions using Github, but that does not seem active so I will just post it here.

The idea would be to add penalties to Robocup Junior Soccer Open (it could be added to Lightweight also, but as far as I know it is used as an introductory to Robocup Junior Soccer).
If a game finishes with both teams scoring the same number of goals teams go into the round of penalties.
There would be four rounds (to save time if needed) and the team that has more goals at the end wins, if they both have the same amount additional rounds are played till there is a team with a higher number of goals, and that team is declared a winner.

Penalty procedure:
The team which defends the goal places black boxes in the goal (same as in the shooter challenge) till place for one box is left.
The ball is placed at the centre neutral spot and the opposing robot at the opposing goal and at the sound of a whistle is turned on. It must not go beyond the centre circle or centre line.

Why I think it is a good idea:

  1. EXPERIENCE
    • Makes the whole experience more fun and like real football
  2. FINANCE:
    • Motors are not that important so it levels the financial playing field
    • No modifications to the playing field are needed
  3. DEVELOPMENT:
    Adds another challenge where students can learn and show their skills
    Adds differentiating factor to Open
    With rules being stagnant for several years now the top and old teams in every country have perfected tactics, this would ensure that even they have to make something new to compete

Why it might be not a good idea to implement it:
1.TIME:
Organizers would need more time to complete the games - this can sometimes be a problem but can easily be solved by removing some time from playing time. If time is not a problem it can be added to every game (the goals are summed)
2. ADDS COMPLEXITY:
While it might be harder than the game by itself the Open should be harder than Lightweight

2 Likes
#2

Hi @Ardi123,

Thanks for your suggestion! I believe it is a very nice one and would be inclined to implement it in the rules!

I have just a few follow up questions:

First, your proposal mentions the following:

Sadly, although the rules currently define the center circle, the center line is not defined. This is a problem for the precision shooter challenge as well, so we’ll have to define it anyway.

That being said, I was wondering what did you mean by not going “beyond the centre circle”? The precision shooter challenge deals with this differently by specifying that

The robot must stay on its half of the field for the goal to count

Did you mean that the robot can cross its own half of the field, if it stays within the center circle?

Second, it seems that without a time limit the “penalty kicks” may take a huge (potentially infinite) amount of time. The precision shooter challenge deals with this by stating that

Each round takes at most 20 seconds.

Do you think something like this would be applicable here as well?

Finally, it is quite unclear to me when would the penalty kicks happen. I am not aware of too many draws that would happen during the RCJ Soccer matches – in fact the rules do not deal with that at all … :slight_smile: Would you mind sharing your reasoning a bit?

I believe it would still be helpful to have it in the rules, it just is not very clear to me where does it come from :slight_smile:

Thanks again for all your interest and activity around here – we really do appreciate that!

#3

Hi Ardi123,
I like this suggestion as well an I think all the technical details (time, center circle, center line) can be solved and are not essential. But I agree with maresuppa, the main subject will be to set the rules when a penalty kick should happen.
The only thing I could image at this very moment would be an essential change of the double defense rule.
Instead of touching the robots to remove one of them out of the goal-zone we could count 5 seconds with double defense
and then give a penalty. This would reduce the referees manual action during the game and would be a real penalty for having a double defense. With this rule teams will automatically try to prevent multiple defense harder then they do it now.
So we improve the gameplay in two sections at the same time.
This rule could be used in lightweight as well, maybe with 10 seconds?
Roland

#4

Hi @stiebel,

Thanks for your reply on this thread!

I very much like your suggestion – it basically makes double defense an equivalent of trying to play with a hand in FIFA Soccer with the penalty being a free kick in both cases.

Pushing this idea further, what I am not clear about in this case is the placing of black boxes in the goal which is what @Ardi123 suggested. In penalties it makes sense as the teams take turns, but how would this work in this “free kick” case? Would a team get say 3 black boxes to place into the goal?

I’d be interested in hearing your thoughts on this.

Thanks!

#5

Well,
a free kick without the opponents robot with the whole goal available would score a point in 99%.
This is to much for my opinion.
Even if we block half of the goal and let the robot start at a ball - touching position would only result in the human calibrating the angle before the robot does a single kick.
So I would say we should block 2/3 of the goal AND the robot has to start in his own half but outside the center circle so that the robot has to move toward the ball and then to kick staying completely inside the center circle. With this arrangement we do not need the center line.

Even more fun could be to let the other team place their robot inside the goal area, but switched off.
That would look much more like a FIFA penalty and would work out fatster than placing special boxes.

We could even think of a tactical time difference.
The “kicker-human” has to start the robot but the robot has to wait for 5 seconds before it starts moving.
during this time, the “goalkeeper - human” can reposition the robot within the goal zone.
But the goalkeeper robot has to bee switched off.
With this arrangement the free kick is not just a “position by hand and kick” but still a
good chance to score a goal. A good robot will change direction to score depending on the goalkeeper robots position.

In this way we would include a real tactical challenge to the game completely without the need of changing anything of the field.

Greetings
Roland

#6

True, it could be changed to “The robot must stay within the center circle”.

Well probability approaches zero so in theory, so there could not be an infinite amount of penalty kicks, but I see your point, it adds a non-deterministic note to the schedule and thus makes organizers’ jobs even more difficult.

True, maybe I was not completely clear, my first idea was to use it to resolve draws, but it could be used in different ways, one example would be if all robots from one team are out of bounds, then the team which does have its robots in the field would be awarded a penalty(this way the penalties would be used as a punishment). The other idea would be to always add four rounds at the end of the game and sum the goals, but I think that @stiebel proposed also very interesting and in my opinion maybe even better rule for handling penalties (to award a penalty after double defense).

#7

Thanks for your responses @stiebel and @Ardi123!

I really like this whole idea, except for the waiting for 5 seconds part.

Here is my counter proposal: the kicking robot would start the “penalty procedure” rotated towards its own goal, inside its own penalty area. Given @Ardi123’s point below, however, it does not seem this would allow too much “premapping” and so those 5 seconds is probably the best thing we can do for now.

To recap this whole discussion, here is how I think the penalties could be administered:

  1. The offending (“kicking”) team places a robot inside its own penalty area, rotated towards its own goal. A ball is placed at the central neutral spot.
  2. The offending (“kicking”) team turns their robot on. The robot needs to stay still for the next 5 seconds.
  3. During these 5 seconds the defending team places a robot which is turned off inside its own penalty area
  4. In order to score a goal the offending team’s robot needs to move the ball inside the opponents goal. It needs to do so in at most 30 seconds and while staying within the center circle once it touches the ball.
  5. If the offending team’s robot moves before the 5 seconds pass, the result is automatically no goal.

Would something like this make sense? We could then apply this penalty to many offences, such as for instance double defense we’ve already discussed. This is a very strong contender for the upcoming rules.

Thanks!

#8

I think that if we do not have any time difference that teams could have three switches, one that shoots center, one left, one right and when the opposition places the robot they could just turn the switch that would from an outside, human perspective be the best at maximizing the probability of scoring and thus remove “the robot deciding where to kick” part of a penalty. This three switch system was an example but various other premapping ideas could be used. We could add that the robot that performs the penalty has to be started the same way every time?

Maybe to punish the goalie for staying in goalie area for long time (pr 10s), since he is really annoying if he only moves inside the goalie area. There could be an exception if the ball is in the goalie area.

1 Like
#9

That’s a very good counter point @Ardi123 – it seems we will have to do counting after all as @stiebel suggested.

This would also be a good addition but it would add a ton of police work which makes a ton of organizational things difficult.

Another option could possibly be an outside signal such as a whistle, but that seems to be too much of a stretch in environments where RCJ Soccer is usually played (i.e. too many games played at the same time).

I’ve updated my recapitulation above – please let me know if something like that would make more sense.

Thanks!

#10

Hi @stiebel!

Thank you for your idea about the penalty. I think it is very interesting.
As an alternative penalty for double defense, I propose to also consider the possibility of a penalty for pushing through. I see from practice that pushing through is more often a double defense.
What do you think about that?

Greetings
Sergei

#11

Hi @Kosachenkosv!
That is an interesting thought.
However, I think we should not have too much penalties.
A penalty for double defense will cause the students to build and program their robots to avoid double defense.
This will lead to a “nicer” game and only if the robots still do a double defense, the penalty shot is a real penalty - much more than just lifting and replacing the robot as it is with the actual rules.
But consider: every penalty shot will stop the game and this should not become normality.
The “pushing-act” is not a violation of any rule, so penalty would not be reasonable.
Only if the ball enters the goal, the “pushing”-rule avoids scoring the goal. This is hard enough I think!
And just practically there is no real way to avoid the “pushing-act” as you get this situation every time a goalkeeper touches the ball and the striker. So pushing is not only the fault of the striker.

However, I would love to debate this and other ideas in Sydney.
Do you take part?

Greetings,
Roland

#12

Hi @stiebel !

I will not be able to participate in Sydney this year. Sorry.
However, your arguments are quite convincing.

Greetings
Sergei

#13

Hi,

I have a imo interesting rule change suggestion.

Change from one camera to allow two cameras.

Reasons:
-currently omni directional mirrors have highly developed by many teams leaving little room for innovation.
-all current cameras OpenCV/ raspberry PI, Pixy, OpenMV etc do not allow the resolution/processing speed for highly accurate field mapping.
-Cameras are relativity cheap compared to other robot components, so cost should not be a major issue.
-Two cameras would potentially allow team to do far more detailed object tracking, including lines and other field markings.
-Allow the ball to be seen at a greater distance especially on the superteam field due to more pixels.
-humans have two eyes.

Cheers
Andrew

2 Likes
#14

Change from one camera to allow two cameras.

Reasons:
-currently omni directional mirrors have highly developed by many teams leaving little room for innovation.
-all current cameras OpenCV/ raspberry PI, Pixy, OpenMV etc do not allow the resolution/processing speed for highly accurate field mapping.
-Cameras are relativity cheap compared to other robot components, so cost should not be a major issue.
-Two cameras would potentially allow team to do far more detailed object tracking, including lines and other field markings.
-Allow the ball to be seen at a greater distance especially on the superteam field due to more pixels.
-humans have two eyes.

I love this idea!
Certainly, two cameras without omni mirror allow more image processing/recognition challenges!

#15

@browny89 I like this idea, too. Especially for new teams, it would be much easier and more competitive for the old teams.

In this thread there is an discussion about the camera rule: