Humanoid Robot HR-2

Denix writes "The HR-2 humanoid robot was constructed during a period of three months at Chalmers University in Sweden. It has 22 degrees of freedom which enables it to easily move around imitating human motions. The robot is also equipped with stereovision giving it possibilities to perform hand-eye coordination. For that task an artificial neural network is evolved. Furthermore, the artificial brain is capable of tracking faces as well as recognising them. The HR-2 is also able to speak. The website also contains a movie (35.5 MB) of the HR-2 in action."

Comments

Awesome!

[MindNumbingOblivion]

Now I know where to go to get my armies of mechanical death.

::CACKLE CACKLE LIGHTNING CACKLE BRRZZT COUGH COUGH::

In seriousness, is there a reason for trying to build a bipedal, humanoid, robot? I mean, this looks cool and all, but what are the advantages (or conversely, disadvantages) to such a design (IANSC [I Am Not Susan Calvin])?

Re:Awesome!

[JanneM]

In seriousness, is there a reason for trying to build a bipedal, humanoid, robot? I mean, this looks cool and all, but what are the advantages (or conversely, disadvantages) to such a design (IANSC [I Am Not Susan Calvin])?

We have buit our entire environment for bipedal movement. Just ask anybody with a walker or wheelchair just how inconvenient (and, without help, occasionally impossible) it is to get around when you're no longer bipedal.

Also, the question of just how we manage to walk, run, climb and so on is interesting in itself. And there's no better way to test ideas than to try them out in reality.

Re:Awesome!

[JanneM]

Real progress will be made only after people stop trying to build a human from machine parts and focus more on usefull applications.

Hard to build useful applications before you know how to make it walk safely, consistently and energy efficiently on at least a reasonable subset of human-constructed terrain types.

I'm always astounded by all the backseat drivers of the world who always know so much better what people should do, without ever feeling the need to do it themselves. Since you want useful applications, go to it. Nothing is stopping you. If you're right that applications is the best focus, you'll undoubtedly eclipse these pitiful, wasteful efforts.

Re:Quantum Consciousness, Not Size, Counts

[HuguesT]

Please note that the Penrose-Hameroff hypothesis is at this stage only an hypothesis, and it most certainly does not explain what consciousness is or how to reproduce it. In fact no one knows if the brain performs quantum computations of any kind, or if such computations are required for consciousness. Some AI luminaries think consciousness is in fact very simple, nothing more than memory.

At this stage physicists are trying to build very simple quantum computers, which could be used for accelerating some very specific computations. Some other people have proven that even quantum computing would not be the panacea that many think it would be. For a start it doesn't give any new insight on how to perform specific calculations that would lead to consciousness. All computations possible on a QC would also be possible on a classical one, albeit usually much slower (but it wouldnt' matter that much, at least in theory).

In other words the Penrose/Hameroff hypothesis doesn't really help in any significant way. It is just saying that the brain somehow performs some magic quantum thingy, and that thingy would somehow be the basis for consciousness. It doesn't say what this thingy is exactly, and most crucially doesn't say how to reproduce it in any way.

the P/H hypothesis is basically just saying "we can't have true AI with the current batch of computers, something else is required", but doesn't say what.

Needless to says this is not very helpful, and might be false entirely.