http://www.nbcnews.com/technology/futureoftech/humanoid-robot-flexes-its-160-muscles-creepy-realism-1C7563685

The bending knee thing looks pretty cool. I wonder how much power they can generate like this.

It’s not clear from the video if this thing can effectively move it’s legs while balancing itself. The walking test seems to indicate that it lacks the ability to correct its balance with its knees bent (or it just has a thing for imitating penguins ). During the knee bending test I could not tell if it was supported in the back or by wires from above.

Let’s face it, walking on two legs is something of a trick. It’s mostly just us and the birds who pull it off well. Especially for tall, lanky humans, there’s plenty of angular momentum to pull us down. We have so many little muscle movements we can make in our feet alone to keep us balanced—something we take for granted. Anyone who’s ever injured their foot and had it hurt to move foot muscles can attest to how quickly walking becomes a challenge.

Instead of trying to build a whole human-sized bot, why not focus on replicating the agility of a human or animal on a smaller scale? Figure out just how many pressure sensors, gyroscopes, and “muscles” it takes to accomplish that before moving on to such a project. This would be so interesting to work on—sometimes I feel like I went into the wrong field.

C R Hunt - Dec 13, 2012:

Instead of trying to build a whole human-sized bot, why not focus on replicating the agility of a human or animal on a smaller scale?

The Japanese seem to be very interested in building very human-like robots as one direction for robotics research. Have you seen the following robots…?

Most Human Like Robot Ever
http://www.youtube.com/watch?v=zIuF5DcsbKU

The Most Human Like Pair of Robots Ever - invented by Japanese engineers
http://www.youtube.com/watch?v=kqt0Yh1ucQE

If you watch enough YouTube videos of robots, you will probably notice that there seem to be several distinct directions of robot research, including: lifelike realism, military walking vehicles, tiny robots that jump or fly for penetration purposes, manufacturing, walking and running on two feet. The robots mentioned earlier would be in the lifelike realism category.

C R Hunt - Dec 13, 2012:

Instead of trying to build a whole human-sized bot, why not focus on replicating the agility of a human or animal on a smaller scale?

Well, for one thing (learned through watching many multiple episodes of Mythbusters), gravity doesn’t scale, so the challenges and difficulties inherent in trying to balance a “full-sized” object don’t exist in small scale, where the relationship between the mass of the object in question, and the energy required to keep it in balance aren’t the same between scales. the smaller you go, and the lower the amount of mass you’re dealing with, the less energy you need, true. But there’s no direct correlation between the two. Additionally, the precision required in small scale is far higher than in full scale, and the smaller you go, the more difficult it is to achieve the necessary level of precision. Sometimes, there is no substitute for full scale testing.