Will the robots be spawned at any place in staging area and we need to detect the entrance?

Issue #112 closed
baila created an issue

Hi all,

I was running my old code. The robots were usually spawned right in front of the tunnel entrance. So I did some blind forward move to enter in the tunnel. Now in Ignition, X1 is spawning further with some orientation as well.

My question is,

  1. Can the robots be spawned anywhere in the staging area?
  2. Do we need to detect the entrance and move towards it?
  3. If Yes, will the entrance shape always be like the virtual_stix.ign map?

Please answer as soon as possible, we need to work on that if that is going to happen.

Official response

  • Alfredo Bencomo
    • changed status to open

    Hi everyone,

    We're looking into moving the artifact origin to the center of the tunnel entrance to be consistent with the Systems Track competition. With this change, the teams will use the /subt/pose_from_artifact_origin service in the staging area to get the robot's location relative to the tunnel entrance. This will simplify the process of entering the tunnel faster at the start on the competition.

    Will this impact or benefit the development your SubT solution? Thumb UP/DOWN?

    Thank you in advance!

Comments (22)

  1. baila reporter

    Thank you very much @Alfredo Bencomo for the answers. One question,

    If I want to spawn the robot at custom position during unit testing, can I do that? If yes, please suggest the steps.

  2. Martin Dlouhy

    ??? Are you serious that the robots should look for the entrance? This is subterranean challenge i.e. the underground where it is dangerous for people, not the staging area? (not mentioning that up to now the team specified the robots location)

  3. Martin Dlouhy

    Why is not the tunnel at location (0, 0, 0), the same as in System Track? Or it maybe is and I just did not get it?

  4. Martin Dlouhy

    Hi Nate, thanks for your answer. I do not see there any information that the robot will be randomly placed? On documentation commit from 24th June I see removal of deprecated Gazebo 9 only.

  5. Joe Palk

    I’d like to ask two clarification questions on this issue:

    1. Is the offset between the artifact origin and the tunnel entrance fixed, or may it change between simulations?
    2. Given the new method of specifying the team of robots, is the pattern in which the robots are initially placed in the world fixed? For example, when the team has four members, is robotName1 always spawned in the forward-left position of the starting configuration?

    Thanks!

  6. Martin Dlouhy

    Alfredo, thanks for the good news (regarding fixed deterministic grid pattern). 🙂

    Is also the tunnel entrance with fixed offset in relation to this starting grid pattern? I.e. if former Gazebo9 version could go simply straight the CloudSim(?) version could use for example turn 30 degrees left and then go straight until it enters the tunnel?

    thanks a lot

  7. Alfredo Bencomo
    • changed status to open

    Hi everyone,

    We're looking into moving the artifact origin to the center of the tunnel entrance to be consistent with the Systems Track competition. With this change, the teams will use the /subt/pose_from_artifact_origin service in the staging area to get the robot's location relative to the tunnel entrance. This will simplify the process of entering the tunnel faster at the start on the competition.

    Will this impact or benefit the development your SubT solution? Thumb UP/DOWN?

    Thank you in advance!

  8. Arthur Schang

    Sounds like a good change to me and reduces some extra headache.

    For those of you looking for some code to start with as far as localizing yourself in relation to the origin, the latest version of wall_follow in the subt_hello_world repository does this to enter a goal relative to the vehicle’s map origin to go to the start location of the tunnel. Currently there is an offset between the entrance to the tunnel and the pose_artifact_origin (and there may be differences between the virtual stix environment and the tunnel practice environments) and this is hard-coded. The source file I’m referring to is located here and the lines relevant to this go from line # 30 to # 83. I hope this helps someone!

    If this change is implemented you can merely assign the goal location to be the pose returned from poseFromArtifact origin and not worry about the static transform between the artifact origin and the tunnel entrance.

  9. GoRobotGo

    Thumbs up. This seems like a reasonable change to better match the systems track. Even as a late change, it seems better than the alternative.

  10. baila reporter

    Great. Of course it is a Thumb UP. Thanks everyone for resolving the issue.

    Cheers. Happy coding to all.

  11. baila reporter

    @Alfredo Bencomo , please post a message here once the service is available in the tunnel_circuit branch.

  12. Log in to comment