Skip to main content: The Technical University of Munich and awtec – ECCEROBOT, the first of its kind
[Translate to English:]

The Technical University of Munich and awtecECCEROBOT, the first of its kind

Seen in this light, it conjures up images of Gunther von Hagens' “Plastination” technique – an ECCEROBOT, one of the first robots with a human-like physique. Artificial bones, muscles and tendons are intended to conceptualize our own bodies and thus stimulate the learning process. A milestone in the development of humanoid robots?

31. July 2012

This ECCEROBOT  (Embodied Cognition in a Compliantly Engineered Robot) is causing a minor sensation. Initially, the upper body with one arm is installed on a platform with wheels. The open torso reveals bones, muscles and tendons. The inner structure and mechanisms of the human musculoskeletal system are replicated. A robot arm with 3D printed parts was also developed as part of this project. Michael Jäntsch (Technical University of Munich) and Alexander Panos (awtec AG für Technologie und Innovation Zürich) were involved in the development of the robot arm, for which 1zu1 Prototypes contributed its prototyping expertise.


In 2000, Honda created the humanoid robot Asimo and achieved a breakthrough. What are the biggest problems when developing humanoid robots?

Michael Jäntsch: The biggest problems in robotics are encountered any time a robot has to move in unstructured environments. With the Asimo, a robot was created which, like an industrial robot, actually still had stiff joints. With this approach, however, you reach the limits relatively quickly. With this ECCEROBOT we've gone one step further and installed flexible drives in the robot. This is, of course, an additional level of complexity.

You've referred to an “anthropomimetic” robot. What does this actually mean?

Michael Jäntsch: The term “anthropomorphic”, i.e. similar to humans, has existed for a long time. With the term “anthropomimetic” we've gone one step further and have also mapped the inner mechanisms of humans: muscles, tendons, the skeleton. So it's about mimicking the system. There's a short movie about our robot on YouTube that shows what I mean.

The premise was „without a body, nothing makes sense“. The robot arm that you built is simple, cost-effective and flexibly reproducible. How did you achieve this?

Alexander Panos: This means that the system can be adapted and reproduced flexibly. This is also different to the previous model, which was still made by hand – its reproducibility was virtually zero.

What were the biggest challenges during development?

Alexander Panos: Ensuring the optimum arrangement of all parts. We had to assemble the robot in a way that enabled us to achieve the required functionality without needing a big suitcase on the side, which would have destroyed our anthropomimetic approach. Originally, the challenge was to house all the components inside the robot itself.

Why did you approach 1zu1 Prototypes?

Alexander Panos: awtec AG für Technologie und Innovation had been working together with 1zu1 Prototypes for a long time. Even the preliminary discussions about the project showed that the technologies in use at 1zu1 could do much more than we realized. In any case, we received advice on an equal footing and were able to base our design on state-of-the-art technology. We are very satisfied with 1zu1 with regard to price and quality. And because we're impressed by their high quality standards, we always go back at 1zu1.

Why is the Technical University of Munich working on robots at all?

Michael Jäntsch: To date there has been relatively little research on humanoid robotics in Europe, and it's only just getting off the ground. Japan is already further ahead in terms of development. The complexity of this tendon-driven robot is especially high and the ECCEROBOT project and a project in Japan are actually the only ones of their kind in the world right now. So the university really is a leader in this segment.

What role did 1zu1 Prototypes play in the development of the robot arm?

Dominik Maccani: Hannes Hämmerle handed over the project to me, I found it incredibly exciting to see what was happening in this field. Our expertise enabled us to produce, join and add hinges to components – which would have had to consist of several elements if manufactured with traditional mechanical-engineering methods – virtually from a „single casting“ using laser sintering technology. We were able to contribute our experience of what is actually feasible today to the design. Our project partners relied on this input and based their designs on our specifications. We thus achieved optimum results.

In mechanical engineering and jig and fixture construction, the designs are almost never aligned with our technological possibilities. However, the ECCEROBOT project shows where the journey could yet take us if we develop the parts together. With learn more and more with each challenging project.
Smart fixture designers are aware of our potential with additive manufacturing – which we are also happy to share with them – and make use of these innovations. In addition, the projects can be realized much more quickly – and at a much lower cost than with the conventional design method.

Robota comes from the Slavic word for „work“ or „slavery“. What application areas do you foresee for ECCEROBOT? Are we talking about service robots?

Michael Jäntsch: So we see this as a multi-pronged approach. From a scientific point of view, a major factor was to understand how humans can move such a complex body. This was one of the main objectives. And also to find out how consciousness arises, how cognition arises. On the other hand, we can imagine certain application areas in the future, such as service robotics, for example, whereby a humanoid robot is placed in the household and empties the dishwasher, does the housework, etc.

Alexander Panos: „In any case, we're not indifferent to this goal and see great potential there, especially at the interfaces of cooperation between humans and robots. Wherever humans and robots need to function together, the work of humans can be made much easier, not only in nursing, but in all possible areas, including safety, for example.

Are scientists in robotics also guided by ethical principles? For example, what if robots go crazy and turn against people, like the Terminator?

Michael Jäntsch: Of course, that's science fiction. We're still a very, very long way from making something like this reality, and if we get there, then our work will only be a very small part of it. Scientists certainly won't be stopped just because there's a possibility that their work might be abused.

What actually motivated you both to get into robotics?

Michael Jäntsch: Coming from the field of computer science, the exciting thing is that you immediately get to see something actually moving. That's what makes it so fascinating. You program it, and then it moves. In addition, the research involves a lot of freedom. The research conditions at the Technical University of Munich are really very good.

Alexander Panos: I come from biomedicine and especially enjoy the diversity of my work. There's always a robotics project going on, and they always constitute an optimal combination of all my knowledge. When it comes to humanoid robotics, I can draw on my medical background. This is very relevant and also the closeness to your own creation and development is simply unsurpassed.

Dominik Maccani, what fascinates you personally about robotics?

Dominik Maccani: The special thing about it is that you're not dealing with the development of an everyday object, like a coffee machine, for example. We're diving into a specialist field – robots in general are fairly common knowledge, but humanoid robots, that's something new. It fascinates me. At 1zu1 we used rapid manufacturing methods to make various parts of the arm, especially the hand, in several versions. The final part was then the whole part. It was laser-sintered in polyamide, unlike the previous model. This allows us to reproduce the parts again and again, directly from the machine, from the data set.

How does the collaboration generally work in such a complex project? Munich, Zurich, Dornbirn. How does the communication work?

Alexander Panos: We use all communication channels, so there are plenty of options. You just have to use them the right way. For example, it doesn't make sense to conduct complex discussions by email, or travel all the way to Munich just to say „Hi“. But choosing exactly the right means of communication for the matter at hand is always a major challenge.

Is the project development continuing with this ECCEROBOT?

Michael Jäntsch: We are definitely continuing our research here. As are the other partners in the ECCEROBOT project. After all, the networks have already been created.

Alexander Panos: With us it's a bit different, because as service providers we always complete our projects. However, the development will certainly continue here as well.

The effects described by Mori can often be read about in forums – a feeling of fear and disgust towards the perfect artificial creature that reminds us of Frankenstein's monster. What about the ambition to create, man playing God?

Michael Jäntsch: So in this project you can already notice that the white, bone-like structures and the tendons are very fascinating, and the question of ethics is always there, but I never really found it repulsive, we've even seen children take the robot in their arms. If the whole thing were to be covered with a skin, then it would start to seem weird to humans. Mori also talked about the so-called „Uncanny Valley“, i.e. the closer you get to humans, the more acceptance the robot gains, but at some point this decreases because it's too close. That's to be expected. Only if you reproduce a person exactly would the acceptance rise again.

Alexander Panos: I agree. As long as you can distinguish it from a human being, the acceptance is high, but if it's almost perfect, then people can't process it so easily anymore, it's like a break with reality. The „obviously wrong“ is more acceptable than the „almost perfect“.

One more question: How do you see the time frame here? When will we have the first human robot?

Michael Jäntsch: Somebody announced the goal of beating a human soccer team with robots by 2050, but that just means the robots will need to be very good at soccer. Building a perfect human robot will certainly take another 50 to 100 years. Alexander Panos: We haven't yet even fully understood how human intelligence works. And until we fully understand this, we can't build a perfect humanoid robot.
 


The questions were put to Michael Jäntsch (Technical University of Munich), Alexander Panos (awtec AG für Technologie und Innovation Zürich) and Dominik Maccani (1zu1 Prototypes) by Hermann Braendle.


Share

Place your order in just a few steps
You can do this with 1zu1direkt

Configure

Technology choice, material, color, surface structure…

Get started now

Data transfer

Upload 3D file, top priority for secrecy

 

Consulting

Analysis, coordination, advice on possible alternatives…

 

Personalized quote

promptly delivered, personalized and perfectly adapted quote

 

Immer Informiert bleiben

Sichern Sie sich einen Wissensvorsprung! Mit dem 1zu1 Newsletter erhalten Sie spannende Neuigkeiten, Informationen zu Messen und Veranstaltungen, zum Produktprogramm und zu besonderen Angeboten.