Professional Documents
Culture Documents
(Meta Package)
Joosoon Lee
2018. 06. 16
1. Introduction
Objective
Background
2. Implementation
Installation
Launch
Motivation
• There are various available services using robots, especially mobile robots.
• For better performance of robot service, the Reinforcement learning(RL)
can be used to train services.
• Simulation is one of the effective and economic way to be used in RL.
Backgound
Reinforcement Learning
: A method by which an agent defined within an environment recognizes the
current state and chooses an action or order of action to maximize reward among
the available actions.
[Agent]
[Action at] [Action at+1]
[Environment]
• Kobuki robot will move in the virtual environment of rviz during simulation.
Implementation
Requirements
• ROS kinetic
• Rviz
• kobuki (meta package)
packages for kobuki driving
available online at (http://wiki.ros.org/kobuki) and (https://github.com/yujinrobot/kobuki)
This manual is for the user who already install kobuki meta package,
who can make turtlebot with keyboard input (kobuki_keyop).
Installation
kobuki_soft (meta package)
(new terminal)
$ roslaunch kobuki_softnode full.launch
(new terminal)
$ rosrun rviz rviz
Add – RobotModel & Odometory
(new terminal)
$ roslaunch kobuki_keyop keyop.launch
Moving with keyboard
Result
$ roscore
Result
$ roslaunch kobuki_softnode full.launch
Result
$ rosrun rviz rviz
Result
Add – RobotModel & Odometory
Result
Add – RobotModel & Odometory
Result
$ roslaunch kobuki_keyop keyop.launch
Result
Moving with keyboard
Discussion