Application for demonstrating surface blending with ROS.

Godel: Austrian logician and mathematician http://en.wikipedia.org/wiki/Kurt_G%C3%B6del

• Install wstool in order manage the repos inside the workspace

  sudo apt-get install python-wstool
  

• Cd into the 'src' directory of your catkin workspace and run the following:

  wstool init . 
  wstool merge https://github.com/swri-robotics/godel/raw/indigo-devel/godel.rosinstall
  wstool update
  rosdep install --from-paths . --ignore-src
  cd ..
  catkin_make
  

• Local mode
• All the functions run at the same machine.
• Network mode
• Functions are distributed to two different machines: Server and Client.
• Server Machine: - Runs perceptions nodes such as surface detection. - Runs motion planning nodes such as the process path planner and free motion planner (moveit).
• Client Machine: - Runs Rviz for user front-end and visualization only. - Runs industrial_robot_simulator node when simulated robot is being used. - Runs robot_driver node when real robot is being used. - Runs generate_point_cloud node when simulated sensor is being used. - Runs Kinect sensor driver node when real sensor is being used.

• Run blending demo in full simulation mode (simulated robot and sensor)

  roslaunch godel_irb2400_support irb2400_blending.launch
  

• Run blending demo in robot simulation mode (simulated robot and real sensor data)

  roslaunch godel_irb2400_support irb2400_blending.launch sim_sensor:=false
  

• Run blending demo in sensor simulation mode (real robot and simulated sensor)

  roslaunch godel_irb2400_support irb2400_blending.launch sim_robot:=false robot_ip:=[robot ip]
  

• Run blending demo in full real mode

  roslaunch godel_irb2400_support irb2400_blending.launch sim_sensor:=false 
  sim_robot:=false robot_ip:=[robot ip]
  

• Bring up Kinect2 sensor, if available:

  roslaunch godel_surface_detection kinect2.launch
  

• Run blending demo in full simulation mode (simulated robot and sensor)

  Server: If it's the first RobotCell instance,run following program:

  roslaunch godel_irb2400_support irb2400_blending_server.launch namespace:=[robot namespace] first_robot:="true"
  

  If it's not the first robotCell instance, run following program:

  roslaunch godel_irb2400_support irb2400_blending_server.launch namespace:=[robot namespace]
  

  Client: Run following program:

  rosrun godel_network irb2400_client.sh [server ip] [robot namespace] false
  

  To automatically run blending cycle (without maually interaction):

  rosrun godel_network irb2400_client.sh [server ip] [robot namespace] true
  

  Note: To support multiple Robots, the server can run multiple ros nodes in the blending application under a unique namespace.

This section contains instructions for performing extrinsic calibration of a camera (on the robot end effector) to the robot arm. This is only necessary if you are running on real hardware with the real sensor.

• Extrinsic calibration routines depend on the industrial_calibration package. Clone this package to your workspace.

• The industrial calibration library builds against libceres, an optimization library, whose installation instructions are available here.