Robots are becoming increasingly similar to humans and may soon be our new helpful housemates. Professor Tamim Asfour is Head of the Department for High Performance Humanoid Technologies (H2T) at Karlsruhe Institute of Technology (KIT) and the »father« of the humanoid robot ARMAR. Our interviewer met the friendly researcher on the KIT open day. She was immediately won over by his infectious enthusiasm for the topic: owning your own robot that cleans and tidies would be truly revolutionary …
Perfect symbiosis? – On the teamwork between robots and humans
Prof. Tamim Asfour, IAR at the KIT, interviewed by Katja Pfeiffer
PROFILE: Mr Asfour, how old is your »child« , the robot ARMAR?
Tamim Asfour: I began to develop ARMAR I and ARMAR II in 1998/99 as part of my doctorate. We presented ARMAR III for the first time in 2006, which you can see in action here. ARMAR IV has just been completed.
PROFILE: Could you describe how you work together with it? Tamim Asfour: Yes, it is of
course a special kind of relationship that you develop over the years with this research and the systems. I have to be fascinated by the research and have a strong affinity with the topic, otherwise I probably wouldn't have stayed here so long! I find robotics research absolutely fascinating, as we are developing systems with humanoid robots that are multifaceted just like humans. That means robots which are capable of carrying out various
tasks: mopping, tidying, cutting, opening doors, using tools etc. Unlike industry robots, they don't just specialise in a single task. Of course, many ask why robots have to look like humans too. The answer is that we don't yet know of a better morphology that is better than or equally as good as the human form when it comes to carrying out tasks in human-centred environments.
PROFILE: So you teach ARMAR human traits .
Tamim Asfour: You could say that, yes. However, I am of course a little cautious with the wording as we are still very much at the beginning. We haven't got far enough to be able to say that these robots can think like humans, for example. Of course we are researching this, but it is still in the early stages.
PROFILE: And when will yo u r robot be an adult?
Tamim Asfour: When it lipreads our requests! And when it understands what it has to do to learn a new specific task. The mature ARMAR will be capable of doing all sorts of things using a set of »innate« abilities, just like us humans. Based on this and on interaction with humans and environments, these abilities develop further. New knowledge is generated to deal with new situations. It's intuitive communication, the ability to learn from observation and from positive and negative experiences. Negative experiences speed up learning significantly. And the better and more memorable the examples are, the more efficient and quicker learning is.
PROFILE: What disciplines do you work with?
Tamim Asfour: Our team is very interdisciplinary and international. We work together with other universities in Germany and abroad. We are computer scientists, mechanical engineers, and electrical engineers, working together with colleagues from cognitive and neurosciences, biology, sports science and material research. In our field, you can really only move forwards when you research across the disciplines. Design also plays a large role here. How do we design a technical system so that it accepts humans? My dream is to be able to go into a furniture shop in a few years and order a kitchen along with a robot that will then do unpleasant household chores. It's here that the architects are needed. Homes today already have to be designed to be easily accessible in part. Robots on wheels need such an environment too.
PROFILE: What could our daily routine look like in future?
Tamim Asfour: The aim is to have a robot like ARMAR in the home that can move, run, clean the table, tidy the children's room, fetch something from the fridge, and much more. But I don't know if it will happen in this way. This technology can of course be formatted in different ways. After all, the systems we are researching and developing here are ultimately just carriers of technology. Even though our aim is to develop complete robot systems like ARMAR, I can also imagine that parts of these systems will soon find use in flexible fabrication and production, and in the future in building services or in domestic appliances,
for example. Why does a table always have to stay in the same place? It could actually be a moving part that knows where space is available, and can move, change its shape, or simply
fold itself away into the wall. What I mean to say is that we are developing technologies that can be incorporated into the designs of homes of the future. Domestic appliances could be equipped with appropriate sensors so that they operate in an anticipatory way. The potential applications of these robotic technologies are limitless.
PROFILE: So a robot assistant that you can relyon.
Tamim Asfour: Exactly! Of course, the question of security is also extremely important here. The robot is working very close to people. You can't allow for any errors.
PROFILE: What is good teamwork for you?
Tamim Asfour: A common understanding of problems and a common language are important. Scientists from different disciplines often have the same problem but talk at cross-purposes. Understanding that precisely what the other person is talking about is also my problem and vice-versa... finding common ground here is no easy process and is perhaps the biggest challenge we are faced with daily.
Humanoid Robotic Systems