The goal of the rest of the chapter is to apply what you have learned about RL in general problems to a specific domain such as robotics. To easily transfer that knowledge, we will reproduce the simple cart pole example, modeling it as a robot in Gazebo. The code samples are in the cart-pole_ROS folder of the code repository of this chapter. Move to that location on your laptop:
$ cd ~/catkin_ws/src/Chapter12_OpenAI_Gym/cart-pole_ROS
Inside, you will find two ROS packages, each one giving its name to the folder:
- cartpole_description contains the Gazebo simulation framework for the cart pole using ROS. The structure of this package is very similar to the one described in Chapter 5, Simulating Robot Behavior with Gazebo. Hence, it is not necessary to dive into its details.
- cartpole_dqn contains the OpenAI Gym environment for the preceding Gazebo simulation...