Tirjen's profile - activity

Hello everyone,

I'm working on exploration of unknown maps with mobile robots. I would like to know if there is a way to update a costmap from a static map (e.g. the one created by gmapping) in place of using the observation_sources parameter, while also keeping the track_unknown_space parameter to true.

I need this because otherwise sometimes my exploration costmap gets noisy. As an alternative can the clear_costmap_recovery be used for a custom costmap in order to reinitialize the map from a static map?

Thank you in advance!

So are the callbacks executing at a lower rate or they are not executing at all? Are the two machines equivalent in terms of CPU, RAM etc?

Can you please post the error that you get during execution?

You can access to PointCloud, Depth image and RGB image. You can use openni_launch to bring up the topics you need.

If the problem is solved, please check the answer as correct

I don't know if you can run the openni_tracker on 2 kinects, but in this case I think the problem is that the second node you run has the same name as the first, which is then stopped. So I suggest you to make a launch file where you launch 2 nodes with different names and different parameters.

http://answers.ros.org/question/91111/rgbdslam-in-ros-hydro/

The undefined reference error means that it can't find the target library. If you haven't done it, you must add the command:

target_link_libraries(test_imu_results ${catkin_LIBRARIES})

Since the roscpp package is in the find_package command I think the problem should be something similar.

P.S.

If I'm not wrong the command catkin_make test_imu compiles all except test_imu package. To compile a single package you should use the command catkin_make --pkg <my_package_name>, but I never tried it.

I suppose you get the error when the add_executable and the target_link_libraries for the test_imu_node are not commented. There is a missing closed bracket in the target_link_libraries for the test_imu_node. I don't know if this is the problem, can you please say otherwise which error gives you the compiler?

Can you please edit your question adding your CMakeLists.txt?

Hello everyone!

I'm trying to install bride following those instructions, but when I try to run the command 'make' inside bride folder I get this error:

mkdir -p build
if [ ! -f eclipse-modeling-juno-SR2-linux-gtk.tar.gz.md5sum ]; then echo "Error: Couldn't find md5sum file eclipse-modeling-juno-SR2-linux-gtk.tar.gz.md5sum" && false; fi
/opt/ros/hydro/share/ros/core/rosbuild/bin/download_checkmd5.py <a href="http://ftp-stud.fht-esslingen.de/pub/Mirrors/eclipse/technology/epp/downloads/release/juno/SR2/eclipse-modeling-juno-SR2-linux-gtk.tar.gz">http://ftp-stud.fht-esslingen.de/pub/Mirrors/eclipse/technology/epp/downloads/release/juno/SR2/eclipse-modeling-juno-SR2-linux-gtk.tar.gz</a> build/eclipse-modeling-juno-SR2-linux-gtk.tar.gz `awk {'print $1'} eclipse-modeling-juno-SR2-linux-gtk.tar.gz.md5sum`
[rosbuild] Downloading <a href="http://ftp-stud.fht-esslingen.de/pub/Mirrors/eclipse/technology/epp/downloads/release/juno/SR2/eclipse-modeling-juno-SR2-linux-gtk.tar.gz">http://ftp-stud.fht-esslingen.de/pub/Mirrors/eclipse/technology/epp/downloads/release/juno/SR2/eclipse-modeling-juno-SR2-linux-gtk.tar.gz</a> to build/eclipse-modeling-juno-SR2-linux-gtk.tar.gz...Done
[rosbuild] Checking md5sum on build/eclipse-modeling-juno-SR2-linux-gtk.tar.gz
[rosbuild] ERROR: md5sum mismatch (027fb35a730ff1f5673a14795a9375f3 != 4d66aa3aeb1c23769fa69f080ccda211) on build/eclipse-modeling-juno-SR2-linux-gtk.tar.gz; aborting
make: *** [build/eclipse/unpacked] Error 1

I triyed both with release 0.2 and branch groovy. I'm working with Ubuntu 12.04 32-bit and with ros hydro. Can someone help me? Thank you!

Thanks I downloaded it using apt-get!

Hello everybody!

I'm trying to use the ORK on ubuntu 12.04 with ros Hydro. I installed it using sudo apt-get install ros-hydro-object-recognition-*. I did not install couchapp (sudo pip install -U couchapp), as, if I understood well, it is not necessary. Now, following this tutorial, I have a kinect connected to my computer and running:

rosrun openni_launch openni.launch

works properly, as I can see the depth image with image_view. However, when I try to use the command:

rosrun object_recognition_capture orb_template -o my_textured_plane

what happens is that the depth image that appears is all black and the ORB window all red. Moreover I noticed that if I try to run this command with the image_view running and showing the depth image, the image_view stops to be updated, i.e. the openni.launch doesn't publish anymore the /camera/depth/image topic. Also when I stop running the orb_template command, the depth image topic still isn't published.

Any help would be appreciated! Thanks!!

EDIT:

this is the output of the command rosrun object_recognition_capture orb_template -o my_textured_plane

Listening for key: s on imshow:save
Registration? 1
Sync? 0
1. device on bus: 2 @ 6 with serial number: A00365A03335051A  Microsoft : Xbox NUI Camera
Time diff max: 16.8667
Registered:on Supported: 1
Setting registration on
Setting sync off

Hello everybody!

I'm trying to use moveit on ros hydro, I'm on Ubuntu 12.04 32 bit. I installed it using the commands:

sudo apt-get install ros-hydro-moveit-full

sudo apt-get install ros-hydro-moveit-full-pr2

when launching roslaunch pr2_moveit_config demo.launch rviz starts well and it seems all ok, but then the MotionPlanning loads something like another pr2 perpendicular to the first and not completely mounted

following the tutorial I changed some parameters like the Planning Scene Topic and finally the second pr2 disappears. But when I move around the start position and the goal position (making sure they are not in collision) and I click the 'plan' button nothing happen. If I try 'plan and execute' the second pr2 appears again and it seems like something is moving, but with the robot disassembled it's difficult to tell it!

Can someone help me? Thanks!!

Hello,

I'm trying to properly use Hector slam in simulation using a turtlebot. I'm using ROS hydro and Gazebo 1.9 on ubuntu 12.04. Once I spawned in Gazebo the turtlebot and the world I created, I use this launch file to start hector slam:

<launch>

  <param name="/use_sim_time" value="true"/>

  <node pkg="rviz" type="rviz" name="rviz" 
    args="-d $(find hector_slam_launch)/rviz_cfg/mapping_demo.rviz"/>

  <include file="$(find hector_mapping)/launch/mapping_default.launch">
    <arg name="odom_frame" value="odom"/>
  </include>

   <arg name="trajectory_source_frame_name" value="scanmatcher_frame"/>
   <arg name="trajectory_update_rate" default="4"/>
   <arg name="trajectory_publish_rate" default="0.25"/>

 <node pkg="hector_trajectory_server" type="hector_trajectory_server" name="hector_trajectory_server" output="screen">
    <param name="target_frame_name" type="string" value="map" />
    <param name="source_frame_name" type="string" value="$(arg trajectory_source_frame_name)" />
    <param name="trajectory_update_rate" type="double" value="$(arg trajectory_update_rate)" />
    <param name="trajectory_publish_rate" type="double" value="$(arg trajectory_publish_rate)" />
  </node>


  <node pkg="hector_geotiff" type="geotiff_node" name="hector_geotiff_node" output="screen" launch-prefix="nice -n 15">
    <remap from="map" to="/dynamic_map" />
    <param name="map_file_path" type="string" value="$(find hector_geotiff)/maps" />
    <param name="map_file_base_name" type="string" value="hector_slam_map" />
    <param name="geotiff_save_period" type="double" value="0" />
    <param name="draw_background_checkerboard" type="bool" value="true" />
    <param name="draw_free_space_grid" type="bool" value="true" />
  </node>

</launch>

The slam seems to work when I navigate the robot around, but I'm very surprised by the bad result compared to gmapping. It get lost very easily and the map is very bad. I would like to know if I set up something in the wrong way, or if it depends from the fact that I'm using a fake laser from kinect or maybe that the world in gazebo I created is challenging for hector slam.

EDIT:

If I move the robot only forward/backward the map is correctly reconstructed, but if I make the robot rotate on itself it prints 'SearchDir angle change too large' and it gets totally lost (sometimes it makes position jumps up to 1 meter and more!!).

This is the world I'm using in gazebo:

Many thanks!!

Hello,

I'm trying to use the rgbdslam package in ROS Hydro (on Ubuntu 12.04), but I'm experiencing some difficulties. First of all I had to install the arm_navigation stack and the octomap_server (I used the hydro-devel catkin version for this one), and I managed to do it. Now the problem is in the rgbdslam package. After following those instructions for groovy:

http://answers.ros.org/question/74899/rosmake-rgbdslam_freiburg-error-in-groovy/ (can't post the right link due to lack of karma),

I'm now trying to build it with rosbuild, but for now it gives me the error:

octomap_ros/OctomapROS.h: No such file or directory

I've seen that this file has been removed in Hydro. Moreover I suppose I will have to change something for the usage of point cloud library, since in hydro has changed very much. Can someone help me to use the rgbdslam in hydro? Thank you!

I think it should work normally... Look for example at this package and see if you are missing something. Here for example JointVelocities.msg depends on JointValue.msg.

Let me know if it works, otherwise please post the CMakeLists.txt you are using.

Please edit also your question posting your CMakeLists.txt.

Again, look at the tutorial I linked in my answer. This assumes obviously that you have done the previous tutorials and that you have programming knowledge in python (or C++).

Please give more details of what you are doing and of the error you get.