https://github.com/deePurrobotics/gazebo_rl

A reinforcement learning environment using Gazebo, ROS and openAI gym
https://github.com/deePurrobotics/gazebo_rl

Last synced: about 2 months ago
JSON representation

A reinforcement learning environment using Gazebo, ROS and openAI gym

• Host: GitHub
• URL: https://github.com/deePurrobotics/gazebo_rl
• Owner: deePurrobotics
• License: mit
• Created: 2018-10-08T19:33:27.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2019-02-07T17:52:18.000Z (over 5 years ago)
• Last Synced: 2024-05-02T18:08:17.084Z (5 months ago)
• Language: Python
• Size: 9.16 MB
• Stars: 15
• Watchers: 1
• Forks: 6
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-Robot-Operating-Syetem - A reinforcement learning environment using Gazebo, ROS and openAI gym

README

        
# Gazebo Reinforcement Learning Environment

This repo contains a few reinforcement learning environements and example scripts greatly inspired by [openai_ros](https://bitbucket.org/theconstructcore/openai_ros.git).

The environments are build with [ROS](www.ros.org) ecosystem, [Gazebo-7.0](http://gazebosim.org/tutorials?tut=install_ubuntu&ver=7.0) simulation software and [OpenAI-gym](https://github.com/openai/gym) toolkit.

## Pre-requisite

- Install [ROS-Kinetic](http://wiki.ros.org/kinetic/Installation/Ubuntu), recommand Desktop-Full install.

- **(Optional but recommend)** Install [catkin-command-line-tools](https://catkin-tools.readthedocs.io/en/latest/)

- **(Optional, skip this step if you've installed ROS in Desktop-Full mode)** Install [Gazebo-7.0](http://gazebosim.org/tutorials?tut=install_ubuntu&ver=7.0)

- Install [OpenAI-gym](https://github.com/openai/gym#installation)

  > `pip install gym`

- ~~**(Optional)** Install [openai_ros](https://bitbucket.org/theconstructcore/openai_ros.git)~~

- Setup turtlebot Gazebo simulation environment


  - Assume you have set up an ros workspace at `/home/yourname/ros_ws`, and ready to config the environment in it. 

  `$ sudo apt install ros-kinetic-turtlebot-gazebo`

  - To make sure `turtlebot_gazebo` is launchable, try `$ roslaunch turtlbot_gazebo turtlebot_world.launch`

- To get ready for reinforcement learning, `$ cd ~/ros_ws/src`, then clone this repo `$ git clone https://github.com/deePurrobotics/gazebo_rl.git`

- `$ cd ~/ros_ws`, then `$ catkin build `

  

## Reinforcement Learning Environments

> All codes are located in `*this_repo*/scripts/`

Copy the model files of cable-driven joint into local gazebo models library

   `$ cp -a *this_repo*/gazebo_models/* ~/.gazebo/models/`

   

### CribNav-v0

1. Launch learning environment, `$ roslaunch gazebo_rl turtlebot_crib.launch` 

2. Open a new terminal, you can `$ rosrun gazebo_rl crib_nav_qtable` to start a q-learning"

   > make sure to `$ chmod +x *this_repo*/scripts/turtlebot/crib_nav_qtable.py`

   

### PlaygPlaygroundFetch-v0

1. Launch simulation, `$ roslaunch gazebo_rl turtlebot_playground.launch` 

2. To test this env, `$ rosrun gazebo_rl pground_env_test`

### CablePoint-v0

1. CMake force plugin for the cable-driven joint

   ```bash

   $ cd *this_repo*/worlds/cable_world/

   $ mkdir build

   $ cd build

   $ cmake ..

   $ make

   ```

2. **Add `export GAZEBO_PLUGIN_PATH=${GAZEBO_PLUGIN_PATH}:*this_repo*/worlds/cable_world/build` to `~/.bashrc`**

3. Launch cable joint simulation, `$ roslaunch gazebo_rl cable_joint.launch`

4. Open a new terminal to test this env, `$ rosrun gazebo_rl cable_env_test`

   > make sure to `$ chmod +x *this_repo*/scripts/cable_joint/cable_env_test.py`