Dr. Erwin Prassler (http://www.inf.fh-bonn-rhein-sieg.de/Prassler.html)

B-IT Bonn-Aachen International Center for Information Technology

and Gesellschaft fur Produktionssysteme

TITLE:

The Robot and the Door Handle

ABSTRACT:

For about 40 years researchers in robotics and artificial intelligence have been working on the design of intelligent embodied agents.

Trailblazing scientific results have been achieved in these forty years. Robotic systems explore the Mars, travel across deserts, move in urban environments, walk and dance, play ping-pong, entertain our kids, and help the elderly.

It seems to be a sign of blissful ignorance and small-mindedness to ask "And what about door handles? Do they know what door handles are?"

Expertise in door handles and door handling is certainly not of earthshaking importance and not key to the future of robotics research either. We still dare to ask to question "What does a robot need to know about door handles and how does this knowledge come about?" More concretely, we investigate the question what knowledge a robot, which at its first activation has only very basic motor skills and perceptual skills available, needs to acquire before it can open a door and leave a room, in which it was locked.

In my presentation I will give an overview of some recent research in the area of robotic discovery and learning by experimentation. I will outline the fundamental scientific question and issues underlying the seemingly trivial irrelevant underlying the use of door handles. I will talk about machine learning paradigms for concept formation and their use for knowledge acquisition by a robotic experimenter. I will talk about grounding of concepts in sensory data and about an evolution of an ontology. I will further talk about robotic curiosity and robotic surprise and motivation for robotic discovery and learning by experimentation. Last but not least I will present some first results on the above issues obtained in a European joint research project XPERO - robot discovery and learning by experimentation.

BIO:

Dr. Erwin Prassler received a master's degree in Computer Science from the Technical University of Munich in 1985 and a Ph.D. in Computer Science from the University of Ulm in March 1996. For his doctoral dissertation he received the AKI dissertation award in September 1997. Between 1986 and 1989, Dr. Prassler held positions as a member of the scientific staff at the Technical University of Munich and as a guest researcher in the Computer Science Department at the University of Toronto. In fall 1989, he joined the Research Institute for Applied Knowledge Processing (FAW) in Ulm, where he headed a research group working in the field of mobile robots and service robotics between 1994 and 2003. In 1999, Dr. Prassler entered a joint affiliation with Gesellschaft fur Produktionssysteme (GPS) in Stuttgart, where has been directing the department for Project Management and Technology Transfer. In this function, Dr. Prassler coordinated the MORPHA project (Interaction and Communication between Humans and Intelligent Robot Assistants, www.morpha.de) one of six national research projects in the field of Human Machine Interaction funded by the German Ministry for Education and Research. In March 2004, Dr. Prassler was appointed as an Associate Professor at the Bonn-Aachen International Center for Information Technology, which is a joint institution between the Univ. of Bonn, the Univ. of Applied Science in Bonn-Rhein-Sieg and RWTH Aachen. He is field of interest include service robotics, in particular technology enabling affordable service robotics, and robotic discovery and learning. He is currently coordinating (another) national research project DESIRE - German Service Robotics Initiative and a European research project XPERO - Robotic Discovery and Learning by Experimentation.