Installation

Dependencies (to be installed separately, before compiling the project)

• MoveIt! (sudo apt install ros-kinetic-moveit*)
• Python (tested with Python 3.5.2 + 2.7.12 installed, working with Python 2.7.12)
• SciKit 0.18.1: sudo easy_install -U scikit-learn==0.18.1 (last version compatible with both Python 2.7 and 3.4, also possible to install it via pip) + sudo apt install python-sklearn
• OpenCV (tested with OpenCV 3.3.1-dev)
• Intel RealSense SDK 2.0 (tested with 2.23.0 version, newer versions may fail)
• RTABMap 0.19.3 (older or newer versions may fail)

Create the CollaborativeBaxter ROS Workspace

• Build with catkin_make (as catkin build will fail because of the ork_recognition package for table top detection)

• Add source /home/YOUR_USERNAME/CollaborativeBaxter_ws/devel/setup.bash to your .bashrc file

• Edit baxter_ridgeback.sh file --> YOUR_IP (check ifconfig output)

• Set ROS_MASTER_URI and ROS_IP in the .bashrc file, or use the sh script if you want to temporary setup these environment variables.

  • # Config for Baxter-on-Ridgeback
    export ROS_MASTER_URI=http://011509P0021.local:11311
    export ROS_IP=192.168.125.48
    

  • Please use the ifconfig command to check what is your ROS_IP (here works ONLY for the INTEL NUC, may be different on your system).

• All the paths are defined to work with the on-board INTEL NUC on Ridgeback. If you need to launch the scripts from your own computer, the easiest way is to globally change the path, thanks to the Global Search tool from VS Code or any editor:

  • Search and replace /ridgebackbaxter/ by /YOUR_USER_NAME/