DontPanic's profile - activity

I am using openni and the Microsoft Kinect so the timestamp is the same. But the compressed depth images still seem to be too big for my wireless.

It is not, that they are really unsynchronised. The time is synchron, but my wireless is the bottleneck. More camera_info messages than actual image messages pass through. So somehow I have to use the ones where two messages are available and ignore the rest. Can message_filters do this?

Hi, I am trying to send compressed depthimages over "/camera/depth/image_raw" topic using image_transport on Wlan. But even compressed they seem to be too big for the Wlan. The program subscribing the topic throws the warning "Topics 'camera/depth/image_raw/compressedDepth' and 'camera/depth/camera_info' do not appear to be synchronized. In the last 10s: Image messages received: 142 CameraInfo messages received: 174 Sychronized pairs: 21"

How can I sychronize the messages. Is there a way to apply another time policy in image_topic without using message_filters or would using message_filters be the way to go?

I tried to split a pointcloud into segments by using PCL in fuerte but, for example, the file "min_cut_segmentation.h" is missing. I tried it with ros-fuerte-perception-pcl-fuerte-unstable getting the same problem.

Is there a way to use those segmentations from PCL within ROS ?

######### Edit 01:

There is no error message in this case. The include files for (most) PCL-segmentation are just not there within ROS, but they are in the standalone version of PCL. As far as I can tell there seems to be no way to use those files out of the standalone version within ROS.

######## Edit 02:

http://svn.pointclouds.org/ros/branches/fuerte/perception_pcl_unstable/

Now it works

Definitely

Trying to compile ar_bounding_box in order to use roboearth I get the following error:

... roboearth/latest/object_scanning/ar_bounding_box src/ar_kinect_extractobject.cpp:43:34: fatal error: pcl/common/transform.h: No such file or directory

Does anyone know how to fix this ?

Great, thank you...

But there a still some problems left: How do you convert the path (type: nav_msgs/Path publisher: footstep_manager/path) into the single steps (type: humanoid_nav_msgs/StepTarget subscriber: cmd_step) ?

And how does it work to really walk the path? Nao seems to be pretty inaccurate when it is about moving. How is it possible to compensate this? In the video it seems to be too accurate to be just calculated prior to the start...

Hey,

I'm just learning how to use Ros for a Nao. I followed the tutorial about how to get the footsteps planned but how is it possible to let the Nao walk this path ? Is there any implementation within the humanoid_navigation or do I need to use the normal navigation stack to work with the published topics from the footstep_planner ?

jbohren wrote:

It looks like you can use the nao_walker.py script from the nao_driver package just by sending humanoid_nav_msgs/StepTarget messages on the cmd_step topic.

@jbohren: Great, thank you...

But there are still some problems left: How do you convert the path (type: nav_msgs/Path publisher: footstep_manager/path) into the single steps (type: humanoid_nav_msgs/StepTarget subscriber: cmd_step) ?

And how does it work to really walk the path? Nao seems to be pretty inaccurate when it is about moving. How is it possible to compensate this? In the video it seems to be too accurate to be just calculated prior to the start...