[ros-users] robot_state_publisher question

[David Lu!!]

I think that would cover my use cases. I guess it depends if we're
talking about full subtrees, or some arbitrary subset of nodes.

If its subtrees, then you could publish the two arms of the PR2
separately, which seems good.However, I have one URDF that represents
our B21r AND the pan tilt unit on top, with the joint states are
coming from two separate nodes. The PTU location depends on the joint
state of the B21r.

I think if its the arbitrary subsets (which I think is what you're
describing), then yes, both my use cases would fit.

-DL!!


On Fri, Oct 1, 2010 at 12:10 PM, Tully Foote <tfoote at willowgarage.com> wrote:
> Ben and David,
> robot_state_publisher does not have persistence currently because if you
> have persistence you must have policies about how long to persist things
> before erroring etc.  These persistence models are very robot and
> application specific.
>
> If you need persistence for a specific application it actually makes sense
> to keep that knowledge outside of the robot state publisher in something
> like David's aggregator.
>
> in general republishing old data is not very good thing to do.  That
> information has already been sent out and has been recorded in tf buffers
> around the system.
>
> One thing which I think might help your use cases is to allow publishing of
> partial trees not starting at the root.  Such that if you provide data about
> M of N  links it will publish data about the M links even if they are
> disjoint and skip the remaining links.
>
> Would this handle your use cases?
> Tully
>
> On Fri, Oct 1, 2010 at 6:35 AM, David Lu!! <davidvlu at gmail.com> wrote:
>>
>> This is a problem I've run into as well. The way I usually fix it is
>> to write a node that aggregates the joint state messages and send out
>> a new stream with all the joints. But I agree, robot_state_publisher's
>> lack of memory is a problem sometimes.
>> -DL!!
>>
>> On Fri, Oct 1, 2010 at 8:07 AM, Axelrod, Benjamin <baxelrod at irobot.com>
>> wrote:
>> > I am using the robot_state_publisher to convert my joint angles to tf
>> > transforms.  However, it seems that the robot_state_publisher doesn’t
>> > like
>> > it when you send it different joints at different times.  In other
>> > words,
>> > robot_state_publisher doesn’t maintain any state…
>> >
>> >
>> >
>> > For example, my urdf has 2 joints:
>> >
>> > Base -> joint1 -> link1 -> joint2 -> link2
>> >
>> >
>> >
>> > If I send it joint1, then at a later time I send it joint 2, I get this
>> > error:
>> >
>> >
>> >
>> > Node: /robot_state_publisher
>> >
>> > Time: 1285876513.251897000
>> >
>> > Severity: Error
>> >
>> > Location:
>> >
>> > /home/arm_user/darpa_arm_software/stacks/robot_model/robot_state_publisher/src/robot_state_publisher.cpp:RobotStatePublisher::publishTransforms:78
>> >
>> > Published Topics: /rosout, /tf
>> >
>> > Could not compute link poses. The tree or the state is invalid.
>> >
>> >
>> >
>> > But if I send both joints at once, then everything is fine.  This seems
>> > like
>> > a major limitation to me.  Am I missing something?
>> >
>> >
>> >
>> > I am using boxturtle.  And am using robot_state_publisher running as a
>> > node.
>> >
>> >
>> >
>> > Thanks,
>> >
>> > -Ben
>> >
>> >
>> >
>> > Ben Axelrod
>> >
>> > Research Scientist
>> >
>> > iRobot Corporation
>> >
>> > 8 Crosby Drive, Mail Stop 8-1
>> >
>> > Bedford, MA 01730
>> >
>> > (781) 430-3315 (Tel)
>> >
>> > (781) 960-2628 (Fax)
>> >
>> > baxelrod at irobot.com
>> >
>> >
>> >
>> > _______________________________________________
>> > 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
>
>
>
> --
> Tully Foote
> Systems Engineer
> Willow Garage, Inc.
> tfoote at willowgarage.com
> (650) 475-2827
>
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>