Discussion of New Rules for 2023

[raymondsheh]

Hi All!

As we discussed earlier this year, we have been wanting to incorporate tasks into the lanes for a while and have demonstrated embedded apparatuses for several years. Based on feedback, it sounds like the time was right to do this now. Of course, the old rules and scoresheets aren’t going to work when teams can do any number of tasks partway through a lap.

There are many different approaches to achieving this, each with different advantages and disadvantages relating to fairness, complexity, flexibility, and ability to achieve a balance between different teams with different, equally important, specializations.

Over the past months we have experimented with and playtested a wide variety of approaches. I’ve put together a one-page form and one-page rule summary that attempts to encapsulate the ideas that have been put forward by myself and the rest of the RoboCupRescue Robot League committee.

The intent is that one piece of paper, form on one side, rule summary on the other, can facilitate the vast majority (possibly all) runs in the competition. There is also a third FAQ page that seeks to provide answers to some of the questions about the choices that were made, but that is not necessary to know how to run a mission.

Of course details of how to actually build and run a lane also exist in separate documents, this is just about how to run and score the mission itself.

Many thanks in particular to Adam for so many new ideas, and for many hours of collaborative playtesting, bouncing around ideas, pointing out problems, suggesting radical solutions, and pushing for simplicity. Thanks also to Ann for looking over earlier versions of this and providing constructive input. This form and rules also incorporates some ideas from rule proposals put forward by Johannes and Martin.

I should stress that this is NOT the final form or set of rules. It is my concept draft - an example of a datapoint within the space of possible rules that seems to provide a balance between all the competing interests in the competition.

This is also not the final list of tests available so if you have suggestions we’d love to hear them! The form even has space for write-ins on both Mobility and Tasks.

I also encourage others who have been considering this to also present their concept drafts, in the form of a scoresheet and rule summary, that we can all playtest and consider over the coming week.

My hope is that we can discuss this week and get input from as many of you as possible so we can consider and, hopefully, come to some sort of final decision, perhaps even in as early as a week or so. So we welcome the constructive input of everyone in the League! The FAQ page, in particular, has discussion points that I am actively seeking input on. Please post here (or, if it’s lengthy, open a new thread) if you have input to provide!

If there is demand, we can also have some online meetings with teams, individually or in groups, to discuss this.

I will also remind everyone that there will be the first RoboCupRescue Robot League roadmap workshop, after the German Open, this coming week. The focus of the workshop is on the direction of the league past 2023, rather than this year’s rules. We encourage everyone who can’t make it in person to join me in participating in the hybrid meetings remotely.

Cheers!

• Raymond

[neng.wang]

Hi Raymond!

I’m Neng Wang from team NuBot-rescue. It appears that there will be significant changes to the competition rules this year based on this summary. We would like to know when the detailed competition rules will be released. Before that, we have some specific questions regarding this rule summary:

1. Will Engage Radio Degradation multiplier be accumulated in autonomous mobility task?

2. Does integrating multiple tasks into lanes mean that the “Best-In-Class” awards will be reduced?

3. According to the rules summary, has the load requirement for the manipulator arm been reduced this year, such as removing the “Strength” task?

4. We do not have a clear understanding of the “Difficult Lane Setting”. For exemple, if we choose the “Incline/Center” lane, does the corresponding “Difficult Lane Setting = Baesline Lane + Negotiate Obstacles” or “Difficult Lane Setting = Baesline Lane + one of the (Negotiate Obstacles, Terrian, Stairs,Crates, Doors)” ? Which specific lanes are referred to as “some lanes have only one option” and what is corresponding option?

5. Are there detailed photos or descriptions available regarding Valves, Buttons and Keys?

6. Will there be separate “Dexterity”, “Search and Inspect” and “Exploration and Mapping” tasks this year? If so, is there an approximate range of height where the objects will be placed in (EXP 3) Recognize Objects?

7. Do we have any information on the format of this year’s finals? We noticed that the preliminary rounds were more geared towards a comprehensive test and were similar to previous finals.

Looking forward to your reply in order to better prepare for the competition.

Best regard!

Neng Wang

[raymondsheh]

Hi Neng Wang,

Yes indeed it will be a significant change!

Note that my proposal is not the final rules and has not been approved by the committee. It is simply a proposal on my end, when I saw that detailed, workable rules were still not yet available from the folks who were pushing to integrate things this year. I wanted to make sure that we had some rules to work with that were consistent with what we announced (rather than falling back onto the old rules for this year). I expect that they will also put forward their proposal, hopefully shortly, so the committee can make some decision and go forward with this. I would hope that this would happen in the next week or so.

Here are the answers to your specific questions, in the context of my proposal. Although it is the only one on the table right now, it has not been approved by the committee and may change.

1: Yes, to make the logistics less complicated, the Radio Degradation multiplier will be accumulated in the Autonomous Mobility task if a team elects to have both. Really this is just so that if a team wishes to flip out of autonomy for a given lap they can still benefit from the Radio Degradation multiplier as we don’t want to deal with the potential randomness that may be caused by trying to flip the Radio Degradation on and off during a mission.

2: I anticipate that the number of Best-in-Class awards will be the same - mobility, autonomy, dexterity.

Of course one of the disadvantages of integrating these tasks is that it becomes more difficult to separate performance - a team has to decide if they are going to optimize on repetitions (mobility) or tasks (dexterity) for instance. Actually this has been a problem in the past with autonomy, where generally a team had to choose if they were going to get points towards Best-in-Class Autonomy and run a lane autonomously, or go for Mobility (and Championship) points and run teleop.

Having said that, teams have the same number of slots that they used to have. So a team that does both mobility and dexterity can easily run a lane focusing on mobility, and then run it again but now focusing on dexterity. I think that there should be some lanes that are easier terrains so that the overhead of needing to do two repetitions before coming back and doing the next dexterity task will not be a huge mobility challenge. So one way to look at this is that, instead of only having one dexterity apparatus, a team can now choose to do dexterity runs in any lane.

We’re open to other ways of solving this issue without going back to the old way of doing everything separately (which we’ve had several teams tell us they don’t want to do). So if you have any ideas on how to solve this problem (or any of these other problems) do let us know!

3: So my original proposal was that the “Extract” would actually be pulling relatively heavy objects out of the arena (e.g. water bottles), rather than the lightweight manipulation objects, thus more objects would mean the robot had to carry more weight out of the arena. Of course it’s much easier for me to push for such a direction if teams tell us that there is demand for tests of weight lifting capabilities. So if you would like this, please let us know!

(Of course also note that Shoring also requires some weight lifting capability as the wooden blocks are also somewhat heavy.)

4: Every lane theoretically has a “Baseline Lane Setting” and a “Difficult Lane Setting”. For example, the “Incline/Center” lane will have the embedded Negotiate obstacles. The problem is that we have practicality issues in the arena. For instance, it is difficult to switch Sand/Gravel from flat to sloped so it is likely to only be offered in one way or the other. Again, I have some ideas for what suitable options here but it’s much easier to make a case if we have feedback from teams. Which lanes would teams particularly like to see offered in “Difficult” setting?

5: I believe this should be coming in a week or so, I know some example apparatuses have been built. In the meantime, here’s a rendering showing the general idea we’re going for.

6: The Dexterity tasks are expected to range from approximately 30 cm to 90 cm above the terrain unless of course teams have a demand for something different. These will be integrated into the lanes, rather than separate as discussed before (but of course it also means you can do dexterity in any lane).

I’m not sure what the plan is for Exploration and Mapping, I’ll let someone else answer that. I would expect that they would be again in the 30-90 cm range.

7: Yes the idea is that we had requests from teams who weren’t making it into the finals to still be able to have a “finals-like” experience rather than just doing ‘boring laps’. In the plan I proposed, the semi-finals and finals will run in exactly the same way and using the scoresheet but, instead of just taking place in a single lane, it will now take place across a sequence of lanes, so when the robot finishes a lane, instead of turning around, it continues into the next lane. Each lane will have a variety of (probably different) tasks, kind of like going into different rooms to do different things. 5 of these together will mean that teams can go up and come back for their 10 repetitions.

So please give us your feedback! We want to build this competition for you but to do that we need to hear from you. We are of course making some big adjustments here based on some rather loud voices saying they want more interesting integrated preliminaries rather than just doing laps, and they want it this year, so we’re doing that, but I also want to make sure that somewhat quieter voices are also heard.

Also please do join the workshop later this week online where we will be discussing the roadmap for the league. I believe details will come from Oskar shortly.

Cheers!

• Raymond

[neng.wang]

Hi! Raymond

Thanks for your reply! We also think that these new competition rules provide a comprehensive test of the abilities of rescue robots, which will certainly make the competition very interesting.

Has the “search and inspect” task been removed this year? Or has it been integrated into the lane as well?

For the load requirement of the manipulator arm, we hope to keep the “Strength” task as a part of “Sensing and Dexterity” because the load capacity of the manipulator arm is also important for rescue robots, for example, to move collapsed obstacles to rescue trapped victims.

Best regard!

Neng Wang

[raymondsheh]

Hi Neng Wang,

I believe “Search and Inspect” is part of labyrinth, although that isn’t a section that I’m working on. I’ll have to leave someone else to answer exactly how that is happening.

I agree that load capacity is important and I’ll push for that.

• Raymond

[inyupbaek]

Hi,
I’m Inyup Baek from team ROBIT.
I wonder when the detailed description of the new rules for 2023 for Radio Degradation or incorporate tasks into the lines will be released.
Best regard!
Inyup Baek

[oehler]

Hello Inyup Baek,
the rule document is scheduled to be released this week.

Best regards,
Martin