First of all we have to clear up what does artificial intelligence means? If we take a regular computer magazine in their opinion is that intelligence that was created in an artificial media with the computer on first place. To have artificial intelligence we first need a model to create it. That model came from the old Mother Nature that has inspired us in many areas. Biology performs some of our most sophisticate computation and exhibits are robustness and adaptability uncharacteristic of our current generations of computers, but will be the characteristic base of the next generations of computers. One way of capitalizing on this realization is to learn and copy techniques from biology and apply them to building and improving our existing computational infrastructure. But we can also consider another agenda -- that of replacing or augmenting the electronics/silicon substrate of modern computation with a living, biochemical substrate. Important engineering advantages include the (unique) capability of self-replication, a straightforward interface to the chemical world, and access to the most sophisticated nanostructural assembly system, the ribosome.
But the achievement in this area proved not to be satisfied yet.
In the eternal battle of creating the perfect robot, which assembles human nature, researchers have created the M2 robot, which is a humanoid that was made for walking research. Since the robot is focused purely on walking it is simply a torso and two legs. It has 12 active degrees of freedom powered by Series Elastic Actuators. In contrast to the other humanoid walking projects in Japan and Europe, control of M2 does not rely on the traditions from robotic arm research, which suggest "stiffer is better". Instead, control of M2 is 'soft' and concentrated on augmenting the natural dynamics of the robot rather than overriding them.
In contrast with the humans...