ros-interactive-manipulation / pr2_object_manipulation Goto Github PK

Seems we have some system calls to the old reconfigure_gui.

• Selecting the "Advanced options" menu on a segmented object cluster fails with the message:

Running system command: rosrun dynamic_reconfigure reconfigure_gui /interactive_marker_node
reconfigure_gui has moved!

 Try: rosrun rqt_reconfigure rqt_reconfigure

Hello,

First of all, thanks for writing this great library!

I'm hoping to

1. catkin-ize this
2. hydro-ize this

But the first step is just to get it to build (with rosbuild and groovy), and I'm seeing:

$ rosdep check pr2_object_manipulation
All system dependencies have been satisified
ERROR[pr2_tabletop_manipulation_launch]: resource not found [mongodb]
ERROR[pr2_interactive_manipulation]: resource not found [common_msgs]

mongodb for ROS doesn't seem to be available anywhere (can't find with apt-get and the repository that the wiki points to is unreachable.

As far as common_msgs go, apt-get says it's already there:

$ sudo apt-get install ros-groovy-common-msgs 
Reading package lists... Done
Building dependency tree       
Reading state information... Done
ros-groovy-common-msgs is already the newest version.
0 upgraded, 0 newly installed, 0 to remove and 181 not upgraded.

I know that this hasn't been worked on in almost a year, but is anyone willing to answer quick questions or update a readme? I'll give a pull request if/when I get a catkin and then hydro branch working.

Add new tools to rviz that publish the 3D click position regardless of what in the scene the click ray intersects. Current use cases:

• Robot, look here
• Teleport gripper here
• Teleport navigation goal here

Note that satisfying all of these might be simply a case of publishing to different topics.

If we end up with too many of these application-specific tools, a solution might be to have them show up / get disabled based on context, though we'll have to discuss how exactly that might look like.

In addition, add an rviz tool for the equivalent of the 'f' button (focus rviz cam).

Hi developers,

is this package still being maintained? If you aren't planning on catkinizing this and bringing this forward to Hydro, I will do so.

-Dash

Problem:

• once the robot sees an obstacle it has a very hard time clearing it from the Octomap
• if the robot moves a tiny bit, laser rays stop hitting particular cells, which never get cleared

As a result, when using for example Gripper Click for grasping in clutter it is often necessary to "Reset collision map" first, a step which we'd like non-expert users to not have to go through.

The first step would be a more careful investigation of the phenomenon to get an idea of exactly what kind of spurious or ghost data is left behind.

Possible short-term investigations and fixes

1. increase resolution of laser

• vertical resolution can be increased by slowing down the laser
• horizontal resolution can be artificially increased by adding one more filter on the laser data (there is already a pipeline in place for that) that supersamples the scan.

2. investigate parameters of existing filters, like the shadow filter, in order to reduce spurious points

Possible solutions which would imply more coding effort

1. increase laser "beam width" explicitly in Octomap by casting more rays

2. add decay for cells in Octomap, likely with purging of old data at serialization time

After #30 is complete, rviz will allow the user to interact with any scene component. At that point, there is no more need for the Interactive Snapshot (IS), which is a bandwidth-hog.

Use the new interaction tools in #30 applied to the Compressed Depth Cloud (CDC) to provide all the functionality currently provided by the IS.

While the IS is low-bandwidth, we'd still like the ability to

• Pause Streaming
• Resume Streaming
• Refresh
  on demand. Build an rviz display plugin that provides this functionality by talking directly to the CDC publisher and sending the commands above. Note that this display will not talk directly to the CDC rendering display within rviz.

Hey there,
I am quite new to ROS and I am trying to use Baxter's head camera to table top image segmentation. I am given a white table with a PCB inside a CPU chassis. I am trying to detect the PCB on the table. Could you please advise how can I your table top segmentation and remodel it to fit a Baxter for my task ?
Any help is appreciated.

RViz related - the Interactive Object Recognition popup crashes RViz with the following message:

terminate called after throwing an instance of 'Ogre::RenderingAPIException'
what(): OGRE EXCEPTION(3:RenderingAPIException): Zero sized texture surface on texture ImageOverlayTexture_1 face 0 mipmap 0. Probably, the GL driver refused to create the texture. in GLTexture::_createSurfaceList at /build/buildd/ogre-1.7.4/RenderSystems/GL/src/OgreGLTexture.cpp (line 406)

As reported by @mateiciocarlie

Hello Matei, sorry to bother you again!

I'm hoping to use this stack for automated grasping, and I see the pr2_pick_and_place_demos is in the groovy-devel branch.

Regrettably, apt-get installs only the groovy version, which doesn't contain the demo. I've built my own demo package with the code, but it depends on pr2_gripper_reactive_approach, which is only in groovy-devel. And per the last thread, it doesn't seem possible build the code.

Do you have binaries available for groovy-devel? I'm on Ubuntu 12.04.

Many thanks for any help,
Max

I'm simply trying to follow the interactive manipulation tutorial to see how interactive markers can be used to control the PR2, but it seems pretty broken.

Ultimately, the problem is that when I run

roslaunch pr2_interactive_manipulation_frontend pr2_interactive_manipulation_desktop.launch sim:=true nav:=true

rviz appears but I'm not able to interact with the robot - no interactive markers appear anywhere.

There are many other hiccups along the way that may be related to this. The fixed frame (/map) does not exist, and I manually change it to /odom_combined (what it would be if the nav argument were false).

Also, after rviz launches, the terminal just cycles the following info over and over...

[ INFO] [1373485727.731924178, 29.689000000]: Waiting for service environment_server/get_robot_state
[ INFO] [1373485727.732991397, 29.689000000]: waitForService: Service [/environment_server/get_robot_state] has not been advertised, waiting...

Can you help me troubleshoot this to get interactive markers working?

Hi maintainers,

would it make more sense to move the pr2_object_manipulation package that exists inside the ros-interactive-manipulation organization over to the pr2 organization?

Dash

RViz added a new functionality to hide non-active controls on Interactive Markers, which unfortunately makes some markers disappear that shouldn't.

Thus, the following controls should have their always_visible flag set to true:

• The virtual gripper in gripper placement mode
• The nav goal robot

@mateiciocarlie are there more?

Selecting "place" when the gripper is holding a segmented or recognized object crashes Rviz with no errors or warnings other than exit code -11. The backend does not crash.

I want to run pr2_interactive_manipulation in a real robot, but it seems pretty broken.

see the output at https://gist.github.com/pxlong/2d3aaeb12e567f7c5547
after I running:
roslaunch pr2_interactive_manipulation pr2_interactive_manipulation_robot.launch nav:=true

see the output at https://gist.github.com/pxlong/eb81e66950854e20be2f
after I running:
roslaunch pr2_interactive_manipulation_frontend pr2_interactive_manipulation_desktop.launch nav:=true

the rviz appears but I can not interact with the robot - no interactive markers appear anywhere.

And the fixed frame (/map) does not exist, and I manually change it to /base_link (I run the both launch file wit nav:=true).

Note that you can see the output of robot.launch, the output just cycles the following info over and over...

[ INFO] [1373485727.731924178, 29.689000000]: Waiting for service environment_server/get_robot_state
[ INFO] [1373485727.732991397, 29.689000000]: waitForService: Service [/environment_server/get_robot_state] has not been advertised, waiting...

Can you help me fix this to get interactive markers working?
Thank you very much.