The robotic legs are designed to mimic human neural architecture, musculoskeletal architecture, and sensory feedback pathways, in order to make the robot move exactly like we do. The robot uses an artificial form of a central pattern generator (CGP)--in humanoids, this is a neural network in the spine that helps us walk in rhythm. An array of different sensors help this robot walk with the same effortlessness as we mere mortals possess. For instance, load sensors pick up on pressure when the robot's foot touches a surface.

While the robot can at present only do a walking gait, it is already helping scientists understand how humans learn to walk. Thanks to the robot, researchers from the University of Arizona were able to conclude that even before babies learn to walk, their bodies already possess a simple CGP, which eventually matures to allow for more complex movements.

Ultimately, these legs will help with scientists better understand spinal cord injuries and, most significantly, what we can do to help paralyzed individuals walk again on their own. For that reason alone, this development could be pretty incredible.

[ via ]