[ros-users] object grasping pipeline and tutorials

Thu Jul 15 23:31:53 UTC 2010

Yes, those two nodes also need to know not to use the database. I'll
think of a better way to put those flags in and put a fix in tomorrow.

Thanks for catching these.

Matei

On Thu, Jul 15, 2010 at 3:42 PM, Hai Nguyen <haidai at gmail.com> wrote:
> I seem to have the same issue.
>
> [ERROR] [1279233614.830963377]: Database connection failed with error
> message: could not translate host name "wgs36" to address: Name or
> service not known
>
> [ERROR] [1279233614.831115827]: Failed to open model database on host
> wgs36, port 5432, user willow with password willow, database models
> [FATAL] [1279233614.831195432]: BREAKPOINT HIT
>        file = /u/haidai/svn/grasping_pipeline/object_manipulation/grasp_planner_database/nodes/grasp_planner_database_node.cpp
>        line=194
>
> [grasp_planner_database-17] process has died [pid 9673, exit code -5].
> log files: /u/haidai/.ros/log/37988d08-9060-11df-b494-001517c01a67/grasp_planner_database-17*.log
> [ INFO] [1279233615.002299804]: Waiting for map ...
>
> On Thu, Jul 15, 2010 at 6:34 PM, Adam Leeper <aleeper at stanford.edu> wrote:
>> Hi Matei,
>>
>> When I run
>> roslaunch pr2_tabletop_manipulation_launch pr2_tabletop_manipulation.launch
>>
>> even though I changed the use_database parameter to false, this error still
>> pops up each time during the launch.
>>
>> [ERROR] [674.702000000]: Database connection failed with error message:
>> could not translate host name "wgs36" to address: Name or service not known
>>
>> [ERROR] [674.702000000]: Failed to open model database on host wgs36, port
>> 5432, user willow with password willow, database models
>> [FATAL] [674.702000000]: BREAKPOINT HIT
>>     file =
>> /home/aleeper/ros/overlays/tabletop_object_perception/tabletop_collision_map_processing/nodes/tabletop_collision_map_processing_node.cpp
>>     line=173
>>
>> [tabletop_collision_map_processing-36] process has died [pid 31557, exit
>> code -5].
>> log files:
>> /home/aleeper/.ros/log/9acfc274-905b-11df-8f09-002354971365/tabletop_collision_map_processing-36*.log
>>
>>
>>
>>
>> And a result, when I try to run the pick_and_place keyboard demo it gets
>> stuck here:
>>
>> [ INFO] [713.141000000]: waitForService: Service
>> [/tabletop_collision_map_processing/tabletop_collision_map_processing] has
>> not been advertised, waiting...
>>
>>
>> Do some of the other nodes also need to know that the database is not being
>> used?
>>
>> Thanks,
>> Adam
>>
>>
>>
>>
>> Adam Leeper
>> Stanford University
>> aleeper at stanford.edu
>> 719.358.3804
>>
>>
>> On Wed, Jul 14, 2010 at 3:52 PM, Matei Ciocarlie <matei at willowgarage.com>
>> wrote:
>>>
>>> Added some fixes to trunk. Thanks for the feedback, it is really
>>> useful as we prepare the (hopefully) stable 0.2 release.
>>>
>>> Adam: there is now a parameter in pr2_tabletop_manipulation.launch:
>>> when starting the tabletop_object_detector node, set use_database to
>>> false and nobody should ever complain about a missing database
>>> connection.
>>>
>>> Hao: the problem you were having was actually a fairly deep bug,
>>> should be resolved now..
>>>
>>> Matei
>>>
>>> On Wed, Jul 14, 2010 at 9:28 AM, Matei Ciocarlie <matei at willowgarage.com>
>>> wrote:
>>> > In order to use the database of models, and thus attempt object
>>> > recognition and execution of precomputed grasps, you need access to a
>>> > database server. Our server at Willow is currently behind the
>>> > firewall, which is why you can not connect to it. We are hoping to fix
>>> > this next week, along with 0.2 release.
>>> >
>>> > For now, there are two options:
>>> > 1)  you can install the database on one of your local machines (or
>>> > even on the robot itself). You can find instructions for installing a
>>> > PostgreSQL server on a ubuntu machine here:
>>> >
>>> >
>>> > http://www.ros.org/wiki/icra_manipulation_demo/Tutorials/Install%20the%20model%20database%20server%20on%20a%20new%20robot
>>> >
>>> > after you are done with the installation of the server, let me know
>>> > and I will provide the database backup file.
>>> >
>>> > 2) do not perform object recognition and grasp objects as unknown
>>> > point clusters. I need to put in a flag somewhere in the detection
>>> > code so that if the database is not present, it will just default to
>>> > this other option instead of crashing. I will do that later today.
>>> >
>>> > Matei
>>> >
>>> > On Wed, Jul 14, 2010 at 9:04 AM, Adam Leeper <aleeper at stanford.edu>
>>> > wrote:
>>> >> I tried this and got these errors. Am I missing some files you have
>>> >> there
>>> >> locally?
>>> >> process[tabletop_node-35]: started with pid [8301]
>>> >> [ERROR] [WallTime: 1279072996.735706295]: Database connection failed
>>> >> with
>>> >> error message: could not connect to server: Connection refused
>>> >> Is the server running on host "localhost" and accepting
>>> >> TCP/IP connections on port 5432?
>>> >> could not connect to server: Connection refused
>>> >> Is the server running on host "localhost" and accepting
>>> >> TCP/IP connections on port 5432?
>>> >> [ERROR] [69.915000000]: Failed to open model database on host
>>> >> localhost,
>>> >> port 5432, user willow with password willow, database models
>>> >> [FATAL] [69.915000000]: BREAKPOINT HIT
>>> >> file =
>>> >>
>>> >> /home/aleeper/ros/overlays/tabletop_object_perception/tabletop_object_detector/src/tabletop_node.cpp
>>> >> line=241
>>> >> [tabletop_node-35] process has died [pid 8301, exit code -5].
>>> >>
>>> >>
>>> >> Adam Leeper
>>> >> Stanford University
>>> >> aleeper at stanford.edu
>>> >> 719.358.3804
>>> >>
>>> >>
>>> >> On Mon, Jul 12, 2010 at 10:28 AM, Matei Ciocarlie
>>> >> <matei at willowgarage.com>
>>> >> wrote:
>>> >>>
>>> >>> Hi Adam,
>>> >>>
>>> >>> The entire grasping pipeline has undergone significant rewriting in
>>> >>> preparation for release. We are hoping to do an 0.2 release of it at
>>> >>> the end of this week, complete with documentation, a stable API, etc.
>>> >>>
>>> >>> In the meantime, if you wanted to be an "early adopter" and check out
>>> >>> the 0.1 version that would be great, as your feedback would help us
>>> >>> for 0.2. You can get the most up to date version using the following
>>> >>> .rosinstall overlay file:
>>> >>>
>>> >>>
>>> >>>
>>> >>> https://code.ros.org/svn/wg-ros-pkg/trunk/rosinstall_files/cturtle/pr2_manipulation-cturtle-pr2all-overlay-devel.rosinstall
>>> >>>
>>> >>> To use an overlay file, you would first do a normal ROS insallation
>>> >>> (of cturtle in this case), then use the overlay file like this:
>>> >>>
>>> >>> rosinstall target_directory overlay_file_path
>>> >>>
>>> >>> And be sure to prepend the target directory to your ROS_PACKAGE_PATH
>>> >>>
>>> >>> To start the manipulation pipeline, use
>>> >>>
>>> >>> roslaunch pr2_tabletop_manipulation_launch
>>> >>> pr2_tabletop_manipulation.launch
>>> >>>
>>> >>> There is no documentation on the wiki yet (we are writing it this
>>> >>> week) but the code itself is well documented. You can get an idea of
>>> >>> how to run it bu looking in the package pr2_pick_and_place_demos,
>>> >>> which contains both Python and C++ examples.
>>> >>>
>>> >>> Let me know if you have any questions, we'll be glad to assist you.
>>> >>>
>>> >>> Matei
>>> >>>
>>> >>> On Sat, Jul 10, 2010 at 2:24 PM, Adam Leeper <aleeper at stanford.edu>
>>> >>> wrote:
>>> >>> > Hi-
>>> >>> > I was trying to follow this tutorial to grab things in
>>> >>> >
>>> >>> >
>>> >>> > simulation: http://www.ros.org/wiki/arm_navigation/Tutorials/Running%20the%20object%20pickup%20pipeline
>>> >>> > but it seems the names of the packages have changed (I realize that
>>> >>> > tutorial
>>> >>> > was last edited February 4th...). Anyway, I found the
>>> >>> > pr2_arm_navigation
>>> >>> > stack but can't seem to find anything directly related to what is
>>> >>> > referenced
>>> >>> > in the tutorial.
>>> >>> > I realize the whole stack is probably still going through a lot of
>>> >>> > changes,
>>> >>> > but do the launch files from that tutorial still exist in some form?
>>> >>> > What is
>>> >>> > the "easy" way of trying out some basic grasping?
>>> >>> > Thanks,
>>> >>> > Adam
>>> >>> >
>>> >>> >
>>> >>> > Adam Leepercoffee
>>> >>> > Stanford University
>>> >>> > aleeper at stanford.edu
>>> >>> > 719.358.3804
>>> >>> >
>>> >>> > _______________________________________________
>>> >>> > ros-users mailing list
>>> >>> > ros-users at code.ros.org
>>> >>> > https://code.ros.org/mailman/listinfo/ros-users
>>> >>> >
>>> >>> >
>>> >>>
>>> >>>
>>> >>>
>>> >>> --
>>> >>> Matei Ciocarlie
>>> >>> Research Scientist
>>> >>> Willow Garage Inc.
>>> >>> 650-475-9780
>>> >>> _______________________________________________
>>> >>> 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
>>> >>
>>> >>
>>> >
>>> >
>>> >
>>> > --
>>> > Matei Ciocarlie
>>> > Research Scientist
>>> > Willow Garage Inc.
>>> > 650-475-9780
>>> >
>>>
>>>
>>>
>>> --
>>> Matei Ciocarlie
>>> Research Scientist
>>> Willow Garage Inc.
>>> 650-475-9780
>>> _______________________________________________
>>> 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
>>
>>
>
>
>
> --
> Hai Nguyen
> CS, Robotics
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>

-- 
Matei Ciocarlie
Research Scientist
Willow Garage Inc.
650-475-9780