-
Notifications
You must be signed in to change notification settings - Fork 61
20220114_tc_meeting
- Time and location: 17:00-18:00 CET over Zoom
- Attendees: Mauricio Matamoros, Matteo Leonetti, Alex Mitrevski, Katarzyna Pasternak
- Note taker: Alex Mitrevski
- Discussion of the outcomes of the survey to the robotics community members
- Simulation league
- General discussion about the rules for 2022
1. There were only 27 responses to the survey. Most of the respondents already have experience with the league; from this point of view, the outcomes are a little disappointing, as the perspectives from current "league outsiders" would have been particularly interesting to see.
2. The competition hasn't yet been confirmed - a confirmation from Toyota is needed - but Mauricio is positive that it will happen. The same rulebook from 2021 will be used (the WRS rulebook).
Edit: A call for participation for the simulation-based competition has been sent in the meantime (on January 17th).
3. General TODOs for the rulebook:
- Replace all instances of 2020 with 2022
- Make sure the language is understandable (use appropriate technical language, no local idioms)
- Ensure that the rules are clear and understandable
- Whenever possible, shorter rules are better
- Check that all of the general rules are applied throughout the tasks, i.e. that task-specific rules don't override the general rules
- Ensure that the evaluation among the tasks is consistent (e.g. picking an object from the same surface should be the same in different tasks)
Concrete discussion about TC suggestions for the rules:
- The general approach the TC aims for is to adjust the scoring of tasks so that easier aspects are worth fewer points and more difficult aspects are worth more, thereby encouraging people to work on the difficult aspects. Is this alright from the ECs point of view?
- Can be done, but the aforementioned consistency should be ensured. Whether this will really encourage working on difficult problems is, however, an open question; it could also happen that teams simply avoid the difficult aspects altogether.
- Can time bonuses be added for tasks if teams demonstrate error recovery behaviours, so that teams are encouraged to work more on reliability aspects?
- No because that would mess up the schedule. This is also difficult for tasks which are anyway long, such as EE-GPSR.
- Can extra points be given for robots facing an operator?
- Might be nice for DSPL, but perhaps it will not be possible for OPL and SSPL, as the robots might not be physically able to do it. This might also be unfair because different platforms have different capabilities. This might make more sense in the context of concrete tasks (e.g. receptionist).
- Looking at the current rulebook as agreed on December 3rd, 2021.
- Alex will start filling out the table in the Roadmap; the contents can then be discussed in one of the next meetings
Friday, January 21th at 17:00 CET (please email the TC for the Zoom meeting link).