Espace ressources

Prerequisite:

• get a running catkin environment.
• check your environment variables.
• check the necessary topics are up and running (cmd_vel, laser, tf,…)
• Filter the LaserScan if necessary.
• Run: roscore and robot drivers and scripts.

Even without having a map of the environment, we can create one on-the-fly, thanks to SLAM (Self Localization and Mapping).

You will need to set up you robot completely, then to launch slam_gmapping. It may be helpful to create a launchfile to pass parameters, for example in gmapping.launch:

  <param name="use_sim_time" value="false"/>\\
  <node pkg="gmapping" type="slam_gmapping" name="slam_gmapping" output="screen">\\
      <!-- descriptions and defaults: \\
          http://www.ros.org/wiki/gmapping#Parameters   -->\\        
      <param name="xmin" value="-5.0"/>\\
      <param name="ymin" value="-5.0"/>\\
      <param name="xmax" value="5.0"/>\\
      <param name="ymax" value="5.0"/>\\
      <param name="maxUrange" value="59.0"/><!--60! ?-->\\
      <param name="maxRange" value="63.0"/>\\
      
      <param name="map_update_interval" value="5.0"/> <!-- def -->\\
      <param name="linearUpdate" value="0.2"/>\\
      <param name="angularUpdate" value="0.15"/>\\
      <param name="particles" value="50"/>\\
      
      <param name="delta" value="0.1"/>\\
  </node>\\

Then, launch rviz and see the result, published in topic /map.

• This metapackage is part of the TU Darmstadt repository.
• To install it in Debian/Ubuntu:

  sudo apt-get install ros-hydro-hector-slam

• For mapping, we're intested in the hector_mapping package.

One may need to first record the data before computing the map. That is possible, with a “bag”.

See the tutorial http://wiki.ros.org/slam_gmapping/Tutorials/MappingFromLoggedData

One first needs to make a bag of data, e.g.: