Package Description

This project provides a ROS stack for teaching a robot through averaging trajectories. This project is based on the bachelor-thesis of Heiko Posenauer and his program Leatra.

Maintainer: Benjamin Reiner reinerbe@hs-weingarten.de University of Applied Sciences Ravensburg-WeingartenInstitute for Artificial Intelligence

Installation

Repository: https://bitbucket.org/BennyRe/ros-lfd-lat/overview

This package requires ROS Fuerte. Other ROS distributions are not tested. Eigen should be installed properly (located in /usr/local/include). First install the Katana Stack and all its required dependencies. Install Orocos KDL with sudo apt-get install ros-fuerte-orocos-kinematics-dynamics (some more dependencies could occur) Now check out this repository and rosmake the ros_lfd_lat package. If he can't find any packages while making, please install them. rosmake the fake_or package, if you need it.

If you want to use the Katana 300 copy the katana package from this repository into your katana_driver directory and overwrite the original one. Then rosmake it.

Running

If you have the Katana connencted over a RS-232-to-USB adapter, then you migth want to run sudo setRightsTty.sh to create a symlink from ttyUSB0 to ttyS0. Start the Katana with roslaunch katana katana.launch rosrun fake or fake or server.py To learn a task: rosrun ros lfd lat lat learner To reproduce a task: rosrun ros lfd lat lat reproducer [trajectory name] The trajectroy name is optional. If you don’t define one the programm asks you about one. Both programs will ask you to do something if they want you to do something.