[ros-users] Gazebo: get_model_state returns wrong relative pose?

Antons Rebguns anton at email.arizona.edu
Tue Oct 5 07:32:35 UTC 2010


John

Right. I will try to come up with a simple world that will hopefully
illustrate the problem we are having.

Thanks,
Anton

On Tue, Oct 5, 2010 at 12:09 AM, John Hsu <johnhsu at willowgarage.com> wrote:

> Hi Anton,
> Can you give a more specific scenario/example?  I.e. what you expect to see
> and what the service call actually returns.
> thanks,
> John
>
>
>
> On Mon, Oct 4, 2010 at 11:51 PM, Antons Rebguns <anton at email.arizona.edu>wrote:
>
>> John,
>>
>> Yes, we did set the relative_entity_name to
>> "robot_description::base_footprint" for our robot coordinate frame. It works
>> as expected if you set that to "world", e.g. the object's pose is returned
>> in the world frame, but it return some unexpected numbers if we set it to
>> robot's footprint frame.
>>
>> Anton
>>
>>
>> On Mon, Oct 4, 2010 at 11:40 PM, John Hsu <johnhsu at willowgarage.com>wrote:
>>
>>> Hi Anton,
>>>
>>> Did you set relative_entity_name to "ModelName::BodyName" of the
>>> reference entity?  Without specifying relative_entity_name the pose returned
>>> will be in the gazebo world frame.
>>>
>>> John
>>>
>>> On Mon, Oct 4, 2010 at 11:25 PM, Antons Rebguns <anton at email.arizona.edu
>>> > wrote:
>>>
>>>> Hello folks,
>>>>
>>>> We hit what looks like a bug in implementation of get_model_state
>>>> service in Gazebo. The scenario that we are working with is the following:
>>>> imagine an object and a robot that is always oriented towards this objects
>>>> and is certain distance away from it. We expect the object's position in
>>>> robot's coordinate frame to be, let's say, 40cm away on x axis and 0 cm on y
>>>> axis. Now if we rotate both the robot and the object by 180 degrees, we
>>>> would expect that relative position would stay the same since robot is still
>>>> looking straight at an object and it didn't move anywhere in relation to the
>>>> object, but gazebo gives us -40cm on x axis and 0 on y axis. These numbers
>>>> get weirder with different yaw angles of the robot with respect to the world
>>>> coordinate frame. So it looks like the absolute location in the world still
>>>> matters when getting the relative pose of one objects in some other
>>>> coordinate frame (except 'world'). Is there something obvious that I am
>>>> missing here?
>>>>
>>>> Thanks,
>>>> Anton
>>>>
>>>> _______________________________________________
>>>> ros-users mailing list
>>>> ros-users at code.ros.org
>>>> https://code.ros.org/mailman/listinfo/ros-users
>>>>
>>>>
>>>
>>> _______________________________________________
>>> ros-users mailing list
>>> ros-users at code.ros.org
>>> https://code.ros.org/mailman/listinfo/ros-users
>>>
>>>
>>
>> _______________________________________________
>> ros-users mailing list
>> ros-users at code.ros.org
>> https://code.ros.org/mailman/listinfo/ros-users
>>
>>
>
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20101005/6ecc3cfd/attachment-0003.html>


More information about the ros-users mailing list