ROS Index
Repository Summary
| Checkout URI | https://github.com/toddhester/rl-texplore-ros-pkg.git |
| VCS Type | git |
| VCS Version | master |
| Last Updated | 2015-09-06 |
| Dev Status | UNMAINTAINED |
| CI status | No Continuous Integration |
| Released | UNRELEASED |
| Tags | No category tags. |
| Contributing |
Help Wanted (0)
Good First Issues (0) Pull Requests to Review (0) |
Packages
| Name | Version |
|---|---|
| rl_agent | 0.0.0 |
| rl_common | 0.0.0 |
| rl_env | 0.0.0 |
| rl_experiment | 0.0.0 |
| rl_msgs | 0.0.0 |
README
```
rl-texplore-ros-pkg
Reinforcement Learning Framework and Repository
Todd Hester
4 May 2011
July 2015 Update
The indigo-devel branch of this repository contains a catkinized
version of the original codebase. If you are using ROS Groovy or
newer, you should check out this version.
The stacks directory contains a ROS stack for reinforcement learning (RL). The code in this repository provides agents, environments, and multiple ways for them to communicate (through ROS messages, or by including the agent and environment libraries). There are 5 packages in the repository:
rl_common: Some files that are common to both agents and environments. rl_msgs: Definitions of ROS messages for agents and envs to communicate. rl_agent: A library of some RL agents including Q-Learning and TEXPLORE. rl_env: A library of some RL environments such as Taxi and Fuel World. rl_experiment: Code to run some RL experiments without ROS message passing.
Documentation for these packages is here:
http://www.ros.org/wiki/reinforcement_learning
And there is a more thorough tutorial here:
http://www.ros.org/wiki/reinforcement_learning/Tutorials/Reinforcement%20Learning%20Tutorial
Working with the ROS build system is difficult at first. Expect a steep learning curve. Working through some of the ROS tutorials is definitely worthwhile when starting out. Documentation:
http://www.ros.org/wiki/ROS.
The interfaces themselves are quite good. So are the visualization and debugging tools. And, there is an impressive amount of useful robotics software already available in the ROS package system.
Directions:
(1) Check out a copy of the RL code base with ROS interfaces
Our packages have their own Github project. Check
out the trunk of rl-texplore-ros-pkg:
$ git clone https://github.com/toddhester/rl-texplore-ros-pkg.git
This creates a copy of the source tree without commit access.
To commit changes yourself, create a github account and e-mail
the project owners requesting commit access.
(2) Install ROS as in http://www.ros.org/wiki/ROS/Installation
This is time-consuming, but not especially difficult. Be careful
to follow the directions exactly. It works fine on Ubuntu and
can be made to work for Mac OSX with some effort. Windows will
probably not work, but there are future plans to support it.
Add the ROS environment setup to the end of your .bashrc:
$ echo "source ~/ros/setup.sh" >> ~/.bashrc
Append our packages as well as the ROS ones (this example does it
from the shell, or you can use any editor):
$ cat <<'EOF' >> ~/.bashrc
export ROS_PACKAGE_PATH=${ROS_PACKAGE_PATH}:~/svn/rl-texplore-ros-pkg
EOF
(3) Log out and log back in.
(4) Try to run some ROS commands:
$ rospack find rl_agent
$ roscd rl_agent
$ rosmake rl_agent
This should compile cleanly.
(5) Try to run an RL experiment. First compile the code:
$ rosmake rl_agent
$ rosmake rl_env
Now we want to run the code to run our experiment. We need to start
both an agent and an environment, and they will interact by passing
rl_msg messages back and forth. Open three tabs in your terminal. Run
File truncated at 100 lines see the full file