ODE fdir1 parameter broken #463
Comments
|
Original comment by John Hsu (Bitbucket: hsu, GitHub: hsu). we need to recover the old bit of code to make fdir1 make use of body-centric frame, see this bit of code that was lost in translation. This will lead to better friction behavior rather than the global frame friction pyramid setting when vehicles are changing orientations. |
|
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). |
|
Original comment by John Hsu (Bitbucket: hsu, GitHub: hsu). fixed by pull request #429 also unveils issue #624 |
|
Original comment by Nate Koenig (Bitbucket: Nathan Koenig).
Fixed in pull request #429 |
|
Original comment by Mike Purvis (Bitbucket: mikepurvis). Sorry to resurrect this 2-year-old ticket, but I'm struggling with understanding what frame fdir1 is specified in. With no way to visualize the friction (GA), I'm left basically referencing ODE documentation (cf). Is there a Gazebo doc I'm not seeing which describes what the deal is with friction? |
|
Original comment by Nate Koenig (Bitbucket: Nathan Koenig). Unfortunately there is no gazebo doc. fdir1 is mostly an ODE thing. Here is where we adjust it |
|
Original comment by Nate Koenig (Bitbucket: Nathan Koenig).
|
|
Original comment by Nate Koenig (Bitbucket: Nathan Koenig).
|
|
Original comment by Jim Hoekstra (Bitbucket: jim3101). I also ran into the problem that fdir1 does not seem to rotate together with the robot, but stays relative to the world frame. Am I missing something or is there still an issue with fdir1? |
|
Original comment by Jeremy White (Bitbucket: knitfoo). I think there is a clear bug in this commit: https://osrf-migration.github.io/gazebo-gh-pages/#!/osrf/gazebo/commits/1ac3c8bb3357 (e75cdf6) The second adjustment sets 'fd2' to the rotated value, but fd2 is never again used. I believe it should, instead, set fd equal to the rotated value of fd2. |
|
Original comment by Jim Hoekstra (Bitbucket: jim3101). Does anyone know if there are any plans for fixing this issue? I would like to use this feature for a project I'm working on. Thanks in advance. |
|
Original comment by Andrea Ruzzenenti (Bitbucket: aerydna). OMG thats's why my simulation is not working anymore! I spent one week trying to understand why friction direction seems to be set in the world reference frame and not in the robot one! Please fix that commit! |
|
Original comment by Mike Purvis (Bitbucket: mikepurvis). Ticket filed: #2068 |
|
Original comment by Jeremy White (Bitbucket: knitfoo). Just a side note, as I've explored this (trying to get a 4 wheeled FRC robot to behave), I have never encountered this bug 'in production'. That is, the way the collisions are chosen mean that my wheels are always the first element in the collision code, so the the fd2 code stanza never triggers. I imagine that if something else with an fdir1 was on top of my wheels I might see an actual problem. So I wouldn't jump to the conclusion that this is the key to your issue. |
|
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). I apologize for the delay in responding, but I will respond in #2068 |
Original report (archived issue) by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).
There used to be code that would rotate the fdir1 vector into the local coordinate frame of the geometry object, but this code is no longer active. This needs to be found and fixed.
The text was updated successfully, but these errors were encountered: