Skip to content

Tutorial 5: Autonomous Mapping

Jump to bottom
samaahitabelavadi edited this page Jul 5, 2017 · 16 revisions

Indoor Environment Mapping Tutorials

Autonomous Mapping

  • Required ROS Packages

    • For using Realsense R200 camera in Gazebo
      1. realsense_gazebo_plugin [2] - [1]
    • For registering the depth image to colour image frame
      1. depth_image_proc - [1]
      2. image_common - [1]
    • For generation of point clouds from colour and registered depth image
      1. rtabmap_ros - sudo apt-get install ros-indigo-rtabmap-ros
    • For frontier based exploration
      1. autonomous_exploration - [1]
    • For path planning
      1. move_it - sudo apt-get install ros-indigo-moveit
    • For keyboard control of the ARDrone
      1. cvg_sim_gazebo - [1]

  • Required Files

    • ardrone_realsense.launch (realsense_gazebo_plugin/launch) - Launch Gazebo7 and load the simulated world
    • register.launch (realsense_gazebo_plugin/launch) - Launch a nodelet to register the depth image to the colour image frame
    • pub_camera_info.py (realsense_gazebo_plugin/scripts) - Run a script to publish fake depth and color camera info for the simulated cameras
    • rtabmap_pcl.launch (realsense_gazebo_plugin/launch) - Launch nodelet to convert depth and color image to pointcloud
    • ardrone_get_odometry.py (cvg_sim_gazebo/scripts) - Run a script to fetch pose of the ardrone in Gazebo and publish the tf
    • moveit.launch (roslaunch move_it/launch) - Launch the MoveIt! path planner
    • send_goal.py (realsense_gazebo_plugin/scripts) - Start the script to send goals to MoveIt!
    • server.py (realsense_gazebo_plugin/scripts) - Start the actionlib server to execute the waypoints from MoveIt!
    • client.py (realsense_gazebo_plugin/scripts) - Start the actionlib client to execute the waypoints from MoveIt!
    • keyboard.py (cvg_sim_gazebo/scripts) - Start the keyboard tele-op to control the drone in Gazebo
    • fbet.launch (realsense_gazebo_plugin/launch) - Launch the node to generate goals for autonomous mapping

  • Procedure

    • Launch ardrone_realsense.launch present in realsense_gazebo_plugin package to start the simulated world along which has the ARDrone with a Reaalsense R200 camera mounted on it
    • Launch register.launch present in realsense_gazebo_plugin package to register the depth image stream to the colour image stream
    • Run the script pub_camera_info.py to publish fake camera metadata for the simulated Realsense R200 camera
    • Launch rtabmap_pcl.launch present in realsense_gazebo_plugin package to generate point clouds from the depth and colour images
    • Run the script ardrone_get_odometry.py present in cvg_sim_gazebo package to fetch pose of the ardrone in Gazebo and to publish the corresponding transform (tf)
    • Launch moveit.launch present in move_it package to start the MoveIt! path planner
    • Run the scripts server.py and client.py present in realsense_gazebo_plugin package to start the actionlib server and client to execute the waypoints from MoveIt!
    • Launch fbet.launch present in realsense_gazebo_plugin package to generate goals for autonomous mapping
    • Start the ARDrone in simulation using the keyboard tele-op script (keyboard.py) present in cvg_sim_gazebo package

  • Appendix

  1. Github Link for packages - https://github.com/eYSIP-2017/eYSIP-2017_Indoor-Environments-Mapping-using-UAV
  2. Link to install Gazebo7 - https://github.com/eYSIP-2017/eYSIP-2017_Indoor-Environments-Mapping-using-UAV/blob/master/bash_scripts/install_gazebo7.sh
  3. Bash Script to launch all nodes and scripts for autonomous mapping - https://github.com/eYSIP-2017/eYSIP-2017_Indoor-Environments-Mapping-using-UAV/blob/master/bash_scripts/autonomous_mapping.sh
  4. Full video link - https://youtu.be/kXyV3OpbWo8
  5. Video for explanation of the autonomous mapping algorithm - https://youtu.be/Ow4pZlDPhkY

Clone this wiki locally