Homework 1 - Perception

Principal aspects

• Rviz configuration for a better visualization of the work environment
• Listener for apriltags on topic /tag_detections to obtain at the same time, for each object in the Kinect FOV, tag’s:
  • ID
  • Size
  • Pose with respect to camera_rgb_optical_frame
• Transformation of the pose’s coordinates to have them with respect to the world reference frame
• Creation of two publisher that expose, w.r.t. world frame:
  1. the requested tags that have to be grabbed with future modules
  2. the not requested tags which will become collision objects to avoid
• Output of these informations in output.txt (inside module folder) as requested in the assignment.

Commands (for simulation)

In separate terminals launch:

• challenge_arena in Gazebo:

    roslaunch challenge_arena challenge.launch sim:=true
  

• apriltags:

    roslaunch challenge_arena apriltag.launch
  

• Rviz to see the results (optional):

    rosrun rviz rviz -d `rospack find g01_perception`/rviz/our_config.rviz
  

• Perception module:

    roslaunch g01_perception discover.launch 
        [ids:="[[frame_id],]"] [sim:=true/false] [forever:=true/false]
  

  • ids is meant to keep the list of wanted tags (separated by ,) that we want to detect: if found, they will be saved in the output file and published within the related topic. The eventually remaining found tags will be published within the other topic as explained above. If the parameter is not present or is empty, all tags will be flagged as wanted. Each invalid tag will cause an error message.
  • sim parameter is used to distinguish between simulated and real environments, mainly for offset-related tuning values.
  • forever is used to keep the perception module running, continuosly publishing its topics.

Additional notes

• Default values:

  • sim:=true
  • forever:=false
  • ids:=""

• If invalid values will be given to sim or forever, default behaviour is used.