evana13g / rapdr_babble

RAPDR Parameter variation capability for CPS

RAPDR_babble

Research in Action Primitive Discovery in Robotics through Parameter Variation

Setup Instructions

Setup each of the following, in order:

1. Ubuntu 16.04

2. ROS Kinetic (http://wiki.ros.org/kinetic/Installation/Ubuntu & http://wiki.ros.org/catkin/Tutorials/create_a_workspace) SKIP the line 'sudo apt install python-rosdep python-rosinstall python-rosinstall-generator python-wstool build-essential

3. Gazebo 7 (Installs with ROS installation)

4. Baxter Gazebo Simulation (http://sdk.rethinkrobotics.com/wiki/Simulator_Installation) a. Note, we are not using the baxter.sh sim script

5. Clone the RAPDR project

    git clone https://github.com/Evana13G/RAPDR_babble.git
   

6. Run the setup.sh script in RAPDR_babble (Only need to run once after installation)

    cd RAPDR_babble && ./setup.sh
   

7. Build:

    cd ~/catkin_ws
    catkin_make
    source devel/setup.bash
   

Run instructions

To run the code, use the following:

Each of the following should be run in a separate terminal window:

1. Launch baxter sim:

    roslaunch baxter_gazebo baxter_world.launch
   

2. Spawn RAPDR Nodes:

    roslaunch agent RAPDR.launch
   

3. Run an experiment:

    rosservice call /experiments_srv [tab complete for args]
        "experiment_name: 'experiment'
         demo_mode: false
         num_discover_strike_runs: 1
         num_cook_runs: 0" 
   

Name the experiment (experiment_name), which RAPDR will use to name the results in the results directory after running. Specify if you would like a demo of the agent accomplishing its goal in the 'original' scenario (demo_mode : true), otherwise the agent will enter the 'novel' scenerio. Specify the number of runs of each type of experiment you would like to run.

4. To run a test script:

    rosservice call /[test name] [tab complete for args]
   

Examples:

    rosservice call /test_actions_srv "{}" [No Args]
    
    rosservice call /test_action_settings_srv "{}" [No Args]

For more instructions on writing/running test nodes, see below which will redirect to the developers wiki.

DEVELOPMENT Run instructions

[FOR DEVELOPMENT MODE] See https://github.com/Evana13G/RAPDR_babble/wiki/Developers-Instructions

Other Info

The URDF models are inside the baxter_simulation package in a folder that I believe is called baxter_sim_examples/models. The URDF model for the table and the wall is called cafe_table.

1. For proof of concept scenario #1, we assume the following protocol:

"left_gripper" - left gripper
"right_gripper" - right gripper
"cup" - cup
"cover" - cover

2. Many of the .py files might not be runnable. Navigate to their directories and run:

    chmod +x [filename]
   

For example:

    chmod +x brain.py