Thomas Ruehr
Thomas Ruehr
Dipl. Inf.
Technische Universität München
Department of Computer Science
Informatik 9
Boltzmannstrasse 3
85748 Garching
Germany
Tel: +49-89-289-17780
Fax: +49-89-289-17757
Office:
Mail: ruehr@in.tum.de
Dipl. Inf.
Department of Computer Science
Informatik 9
Boltzmannstrasse 3
85748 Garching
Germany
Tel: +49-89-289-17780
Fax: +49-89-289-17757
Office:
Mail: ruehr@in.tum.de
My research is focused on robotics in the context of AI. What we call Generalized Pick&Place is moving arbitrary objects in an everyday kitchen environment, where the operation of articulated furniture cabinets is necessay as well as bi-manual grasping. A part of my work, I am aiming to provide robust approaches for such Pick&Place actions that improve their robustness and efficiency over time, using e.g. learning techniques and probabilistic models. For doing so, we also work on the bootstrapping, such as opening an unknown cabinet for the first time, modelling its shape etc. and make such information available in a semantic map.
Working mostly on the PR2 platform, I can offer projects in the area of ROS or the PR2 robot and the Point Cloud Library to ambitious students with or without previous experience. The PR2 is a very forgiving platform with a well documented set of software packages, including tutorials that can be completed in simulation - meaning we can get students to working on a real full-fledged Personal Robot very quickly. Apart from the PR2 we just got two Turtlebots that will be used mostly for student projects and teaching.
@InProceedings{09ETFALearning,
author = {Andreas Leha and Dejan Pangercic and Thomas R\"uhr and Michael Beetz},
title = {Optimization of Simulated Production Process Performance
using Machine Learning},
booktitle = {Proceedings of Emerging Technologies and Factory Automation (ETFA).},
year = {2009},
bib2html_pubtype = {Conference Paper},
bib2html_rescat = {Learning, Planning},
bib2html_groups = {Cogmash},
bib2html_funding = {CoTeSys},
bib2html_domain = {Cognitive Factory},
}