ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

I can't solve your problem, but I can offer some encouragement. Pointcloud_to_laserscan was working very nicely in ROS "unstable" (just prior to Diamondback) from the "ni" stack. It would launch the openni_camera, pointcloud_throttle and kinect_laser nodes. I was able to adjust the "thickness" of the laser to match the height of my robot. It ran well on a robot mounted laptop communicating with a wireless network connection to rviz running on a stationary client computer. Launching was as simple as "roslaunch pointcloud_to_laserscan kinect_laser.launch". Looking at rxconsole, it indicates that all of the nodes/nodelets (openni_camera, pointcloud_throttle, kinect_laser) are passing back and forth nine "openni_camera/bond" messages and the laser displays in rviz in a satisfying manner.

I have not yet been successful in Diamondback. I have the openni_kinect stack and the turtlebot stack which now contains pointcloud_to_laserscan, but "roslaunch pointcloud_to_laserscan kinect_laser.launch" gives an error, "while processing /opt/ros/dturtle/openni_kinect/openni_camera/launch/openni_kinect.launch: "Invalid roslaunch XML syntax: [Errno 2] No such file or directory: u'/opt/ros/dturtle/openni_kinect/openni_camera/launch/openni_kinect.launch' "

I modified kinect_laser.launch to point to openni_node.launch in openni_camera, but the various nodes/nodelets did not link together to pass messages successfully. I am not familiar enough with nodelets to describe what is happening/not happening and have not yet determined how to duplicate my previous success.

Tully Foote is the original author of the pointcloud_to_laserscan package and we can thank him for this useful package. Hopefully he will drop by this question one more time to straighten us out.

The pointcloud_to_laserscan package is probably pretty popular with many of us looking for an inexpensive "laser" to experiment with the navigation stack as you described. It seems to have been bounced around from stack to stack looking for a permanent home. It would be pleasant if this package was added to the openni_kinect stack and the launch file was adjusted for easy success.