As you may have noticed from the various (infrequent and generally late) communications both here on the forum and via other channels, this year’s RoboCup competition will be “Worldwide”, meaning it will be a fully remote event.
That brings its own set of issues but we’d still try to have a couple Technical Challenges, as outlined in the section 9.3 of the rules, even in this remote setup.
Now, the rules already have a couple suggestions, such as precision shooter, penalty kick or vertical kick and we also have a couple more that we’ve thought of internally. That said, we’ve also seen numerous times that the broader RoboCupJunior Soccer community can surprise us significantly with is creativity and is capable of generating ideas we would have never thought of.
Thus, we would like to ask for your suggestions for the technical challenges that you think could be both doable and fun in this remote setup. We generally assume access to a RCJ Soccer field and the equipment that you would assume can be found in a “lab” where RCJ Soccer robots are being built. Furthermore, the challenge should also be something that doesn’t take more than a couple of hours to solve, as we’ll be doing that in 24-hour periods
In particular, we’d greatly appreciate suggestions for challenges that could involve more SuperTeams! I know, it does sound crazy but we’d still like to try it out. Note that the SuperTeams will be done a bit differently this year: in particular, we’ll try to make it so that the teams that form a single SuperTeam are not that far apart from one another in terms of timezones.
As always, should you have any feedback/questions and in particular suggestions, do let us know in here – we’d very happily hear them out
How about freestyle trick shots?
Robots without a kicker or dribbler may not have means of doing fancy shots on a field with goals, but they could be creative and build some sort of paper or wooden guide rail along which the ball could roll, or just think of something? Bonus point if the ball ends up in the goal, but the purpose would be to do cool unexpected shooting tricks with the robots.
Hi, here are some starting ideas for further challenges, if suitable:
one on one attack:
offending robot starts at its goal with a ball.
defending robot is placed 0.75 m in front of its goal and moves only left and right,
trying to trace and follow the position of the ball as best as it can
offending robot must score a goal by first passing the defending robot and then
placing the ball to the goal without violating other usual rules
a goalie challenge
defending robot is placed on a randomly selected neutral spot close to its own goal
a human rolls the ball the fastest speed he or she can from the neutral spot that
is diagonally on the opposite side of the field (close to the opposite goal), making sure
the ball is rolling, not sliding or flying, and along a line that ends in the center of the
target goal.
the defending robot must prevent the goal
attack with passing ball
two robots of the team start on the neutral spots near its own goal, one of them has the ball
while moving generally only forward, they need to pass the ball one to another at least 3 times,
and finally score the goal
slalom
two defending robots are placed on the central axis of the field, one on the intersection of
the central circle with the axis, closer to the offending robot goal, the other one in the middle
of the two neutral points on the defending side
one offending robot starts at its goal and has a ball under the control
the task for the offending robot is pass around the two robots from alternating sides
(i.e. first from the left, and the second from the right, or vice versa), while keeping the
ball under the control all the time, and finally scoring the goal.
catch the ball challenge
an offending robot starts at randomly selected neutral spot near its goal
a human rolls the ball in a slow motion (but fast enough to cross the whole field)
diagonally - from a corner of the marked field on the defending side to the opposite
corner on the offending side (the other corner than the selected neutral spot)
the offending robot notices the ball, catches it, negotiates it, and without further
complications navigates towards the goal and scores without violating usual rules.
I like the freestyle trickshots and one on one attack. I worry that some of the suggested challenges will be hard to compare. Maybe we could add some randomness and give more points for the number of successful attempts out of ten or something.
Here are some ideas of my own:
Scoop
A ball is placed to the right of the field outside of the line. The further outside of the line it is, the more points are given. The robot starts on the neutral point between a goal and the ball. The robot needs to get the ball from outside of the line and score without going fully out of bounds.
Speed scoring
A ball is placed on a neutral point and the robot is placed on the neutral point diagonal from the ball. The robot must score the ball in the least time possible.
Goalie avoidance
A robot sized obstacle is placed in front of the goal in a random location and the attacking robot starts with the ball in the center of the field. The robot must score without touching the goalie. This could be repeated for 5 takes to make it a more competitive challenge.
Control
The robot starts at the center of the field and needs to move to each neutral point facing a previously specified direction. P are scored for speed and accuracy.
Hard stop
The robot starts at a neutral point and must reach the line on the opposite side as quickly as possible. No points are given if the robot goes out of bounds.
One with these challenges is that teams with powerful motors will be favored, but I think this is fine since speed is an essential element of soccer. I also think strong programming can make up for slower robots.
My main idea, to be clear, is to do cool shots on a regular field without additional gimmicks. If the robots aren’t equiped to do so, teams could be creative to use objects that could help them do something cool with the ball. For example, vertical wooden blocks or board could be done to do a pool kick off of a wall. However, the IR ball might not bounce very well. So, instead, one could use a wooden board at an angle to have a hill. If the robot can kick the ball up the board at an angle, it will make a curve and come roll back down. “Guide rail” was meant in this wide open sense, whatever might work for the robots in question - maybe with a preference to keep it simple, I’m not thinking of building a Rube Goldberg machine.
Two teams, each with a field, 2 robots and a ball, do a remote penalty kick competition as follows:
Team A in location A is attacker. Attacking team places one robot and the ball on its half of the field. Defending team (remotely from location B) tells the offending team via video where to place its second robot (power off) in the defending half of the field in location A, to protect the goal. Once defending team is happy with the position of the defending robot, the attacking team activates the attacking robot, in order to shoot a goal. Scoring a goal without touching the defending robot gives 2 points, a goa with touching the defending is 1 point. If ball or robot go out of bounds, no point.
Switch roles, i.e. do the same thing with the robots on the other side of the video call in location B. Take turn until one team has 10 points, or take 10 turns each.
It would be good to have a judge join the video call, to make sure.
To keep everything on the same place, here’s the suggestion that I mentioned before, based on a talk with Ivan:
He suggested a kind of collaboration challenge for the simulation (kind of superteam for two): program two teams to score one goal at a time. Suppose teams A and B are playing. Team A scores, then team B must score, then team A… etc. Each correct score increases the amount of points, while a score in incorrect order decreases it. The (super)teams that score more point in a certain amount of time wins. We can also have one team programming all the robots, but I think it is a nice challenge for superteam.
How about a challenge for Self-position estimation and Ball-position estimation accuracy.
These are the rules.
1.Place 1 robot in the center
2.Put down the ball somewhere random(note. robot shouldn’t chase ball yet)
3.place ball for 5 seconds
4.pick up the ball(note. the robot should identify that the ball was picked up)
5.robot waits for 5 seconds
6.robot moves to the place that the ball should have been
7.measure the difference between the ball that was used to be placed and the place the robot is now.
There is a proposal for a technical challenge. Two years ago at a tournament, there was a technical challenge for lightweights to chase an orange ball. How about shooting the orange ball and the IR ball into two different goals? Here is a gif of the image.
For Soccer Simulation there could be Technical Challenges introducing use of the sensors or alternate WBT files, like:
Everything would work as usual, but the referee would not send translation data for B3. Instead the challenged teams should write a code to use a GPS sensor for B3.
Same as previous but this time the referee would keep orientation data for B3 empty as well. Instead the challenged teams should write a code to use a Compass sensor for B3.
The referee would not send any translation or orientation data for any B team member. Challenged teams should write a code to share their robots’ location and orientation data via communication between their robots. Y team position and orientation data would still be provided by the referee for Y team members.
No Y team member data would be provided by the referee anymore. Instead the challenged teams should write a code to use a LIDAR sensor to have Y team robot positions.
Playing with WBT properties like gravitation, weights of the objects. etc.
Using a bit different shape of robot, capable to lead the ball (not a dribbler, but maybe a curved shape at the front and maybe the end of the box shaped robot or similar approach).
