ROS/Gazebo simulations of sphere robots for reinforcement learning activities and ROS communication for Sphero SPRK+ robots.

Designed in Ubuntu 16.04 using ROS Kinetic and Gazebo 7.

Make sure the following commands have been added to your .bashrc file or ran in the terminal used to launch rosrun, with [INSERT_PATH_TO_REPO] replaced with the actual path to where this repository has been cloned (without square brackets):

source /opt/ros/kinetic/setup.bash
export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:[INSERT_PATH_TO_REPO]/plugins/build
export PYTHONPATH=$PYTHONPATH:[INSERT_PATH_TO_REPO]

Create and configure a virtual environment for managing Python packages:

sudo apt install virtualenv python-opencv
mkdir python2_env
virtualenv python2_env --python=python2.7
source python2_env/bin/activate
pip install numpy pyyaml rospkg catkin_pkg getkey

In a terminal,

cd [REPO_PATH]/plugins/
mkdir build
cd build
cmake ../
make

In a terminal, run roscore

In a second terminal,

cd [REPO_PATH]/arenas/
source /usr/share/gazebo/setup.sh
rosrun gazebo_ros gazebo ctf_1v1_arena.world

In a third terminal, run the first simple proportional control agent script:

source ~/python2_env/bin/activate
cd [repo_path]/agents/
python red_simple.py

In a fourth terminal, run the second simple proportional control agent script:

source ~/python2_env/bin/activate
cd [repo_path]/agents/
python blue_simple.py

In a fifth terminal, run the simulation tracker python script:

source ~/python2_env/bin/activate
cd [repo_path]/host/
python sim_tracker.py

The red and blue spheres will both attempt to reach the other teams 'flag' and return to their own base as quickly as possible.

No longer accurate. Please see Autonomy Hackathon wiki for updated instructions.