Open behradkhadem opened 11 months ago
PS: I think this is an issue:
I think the issue stems from incorrect bounds for end-effector inside URDF file, right?
I think the issue stems from incorrect bounds for end-effector inside URDF file, right?
Probably, have you tried to fix that?
This is issue is the same issue I had with panda with gripper arm (https://github.com/maxspahn/gym_envs_urdf/issues/201) but for mobilePandaWithGripper.urdf. I think I should use MobilePandaFk for this matter, but there is no example or anything for this robot. What is the correct way of getting the position of end-effector in this urdf file? I'm looking for something like https://github.com/maxspahn/gym_envs_urdf/issues/201#issuecomment-1650344088.
The best is to use the generic_fk from the forward kinematics package, then you can directly pass your urdf file.
Probably, have you tried to fix that?
No, first I wanted to make sure that problem stems from the URDF file. If so, I'll have a go at it.
The best is to use the generic_fk from the forward kinematics package, then you can directly pass your urdf file.
Thanks for the suggestion, but what values should I use for child_link
and parent_link
?
Thanks for the suggestion, but what values should I use for child_link and parent_link?
child_link
is the link of the end-effector you are concerned with and parent_link
is usually the root link of the kinematic chain.
child_link
is the link of the end-effector you are concerned with andparent_link
is usually the root link of the kinematic chain.
I understand it, but which one I should use for panda mobile (with gripper)?
https://github.com/maxspahn/gym_envs_urdf/blob/fdfe3d99a8f21c15581930e54d67ed738cf836da/urdfenvs/assets/mobile_panda/mobilePandaWithGripper.urdf#L8-L40
There are three baselinks: base_link_x
, base_link_y
and base_link
. Which one is it?
I don't understand, why are we using GenericUrdfReacher
and not GenericDiffDriveRobot
(like we do with Albert)? Aren't they both non-holonomic? And it's confusing to choose the base link because out base is moving. Or am I misunderstood?
I don't understand, why are we using GenericUrdfReacher and not GenericDiffDriveRobot (like we do with Albert)? Aren't they both non-holonomic? And it's confusing to choose the base link because out base is moving. Or am I misunderstood?
The mobile panda is holonomic.
I understand it, but which one I should use for panda mobile (with gripper)?
Try world
. It should usually work for mobile robots.
Try
world
. It should usually work for mobile robots.
Thanks! What should be the input of fk object? I can get meaningful results by using this:
joint_position = observation['robot_0']['joint_state']['position'][:11]
Is it correct? (fk input can't get the whole observation and says it should be 11-by-1 and not 12-by-one).
And also, I think the issue with grippers getting out of bounds (image above) is not from URDF file, but from pyBullet itself. Because URDF file has the limits (for the fingers) set correctly: https://github.com/maxspahn/gym_envs_urdf/blob/fdfe3d99a8f21c15581930e54d67ed738cf836da/urdfenvs/assets/mobile_panda/mobilePandaWithGripper.urdf#L360C1-L394
Hi @maxspahn ,
This is issue is the same issue I had with panda with gripper arm (https://github.com/maxspahn/gym_envs_urdf/issues/201) but for
mobilePandaWithGripper.urdf
. I think I should useMobilePandaFk
for this matter, but there is no example or anything for this robot. What is the correct way of getting the position of end-effector in this urdf file? I'm looking for something like https://github.com/maxspahn/gym_envs_urdf/issues/201#issuecomment-1650344088.