zeinab's profile - activity

Hi,

I have trouble with a nodelet which is responsible for converting depth image to laser scan. I have the required topics namely /camera/depth/camera_info and /camera/depth/image_raw and here is how my launch file looks like:

<launch>
  <arg name="camera"  value="camera"/>
  <arg name="manager" value="$(arg camera)_nodelet_manager" />

  <group ns="$(arg camera)">
    <node pkg="nodelet" type="nodelet" name="$(arg manager)" respawn="true" args="manager"/>
    <node pkg="nodelet" type="nodelet" name="openni2_camera" args="load openni2_camera/OpenNI2DriverNodelet $(arg manager)" respawn="true">
    </node>

    <node pkg="nodelet" type="nodelet" name="depthimage_to_laserscan_loader" args="load depthimage_to_laserscan/DepthImageToLaserScanNodelet $(arg manager)">
        <param name="scan_height" value="10"/>
        <param name="output_frame_id" value="/depth_camera_link"/>
        <param name="range_min" value="0.3"/>
        <remap from="image" to="/depth/image_raw"/>
        <remap from="scan" to="/scan"/>
    </node>
  </group>
</launch>

My problem is that the /scan topic is empty and is not being published. However when I do

rosrun depthimage_to_laserscan depthimage_to_laserscan image:=/camera/depth/image_raw

the /scan topic is being published as expected. Do you have any suggestions?

Thanks a lot in advance.

importing rospy and actionlib causes embedding python code in c++ to fail Hi, I am using a large code base were our robo

I have found that openpose and its wrappers seem very promising at the moment.

importing rospy and actionlib causes embedding python code in c++ to fail Hi, I am using a large code base were our robo

Hi, Have you found a good solution?

I am not sure if this helps, but I found out that this happens when a new goal is sent. In my case, I was only intereste

Sure, but I thought there might be an element of ROS such as threads that could lead to this, since I could not think of

'Path' object has no attribute 'currentPosition' I have a very strange case here. I have a python node that has a class

'Path' object has no attribute 'currentPosition' I have a very strange case here. I have a python node that has a class

'Path' object has no attribute 'currentPosition' I have a very strange case here. I have a python node that has a class

Please let me know what you would like to see more.

Thanks for your reply. Well, its not copying but rather creating a new object with similar fields for the variables of c

'Path' object has no attribute 'currentPosition' I have a very strange case here. I have a python node that has a class

'Path' object has no attribute 'currentPosition' I have a very strange case here. I have a python node that has a class

I was wondering if there is already an off the shelf solution since many people are using the Kinect for this purpose nowadays.

I need a detector and tracker in combination. I could limit the speed of the robot to an extent. I am more considered about the robustness of the detection-tracking of the people (even at the expense of having to use lower speeds)

Thank for your prompt reply. I am targeting an indoor application. The Kinect is mounted on the robot at the height of approximately 1.2 m and the ground can be assumed to be planar. the robot can move up to 2 .5 m/s.

Hi,

I am looking for a robust person tracking ROS-based solution for the Kinect. I know there exist a number of different packages but I was wondering which one you recommend as a robust tracking solution when using specifically the Kinect.

Thanks a lot in advance.

Hello,

I have a problem regarding obstacle avoidance and local planning. I have a short range Primesense sensor which gives me the scan for my navigation. I am able to do mapping, localization and fairly good navigation when all obstacles are already incorporated in my map. The problem is that dynamic obstacles are detected and shown in my local costmap but the robot does not take them into account somehow. The strange thing for me is that it inflates the obstacles which are in the map but not the dynamic obstacles even though I have inflation_radius set in my local_costmap_params.yaml. I have this suspicion that it might be that the obstacles are only detected 35 cm away from the kinect and otherwise it thinks that the map is clear and since it plans for a short distance in each step it cant detect the obstacles ( I don't know if it makes sense!) I tried changing the sim_granularity: 0.2 but didn't see a difference. Does anyone have an idea about this? I have attached a snapshot of my rviz.

Thanks a lot in advance.

C:\fakepath\Capture.JPG