Robot learns to walk like a human

Humanoid Robots (walking on two legs) currently still have big difficulties with walking on uneven surfaces. American scientists from Human and Machine Cognition Robotics Lab in Pensacola in Florida developed together with Boston Dynamics (a Google subsidiary) a new algorithm which helps the robot to test the surface and take appropriate actions.

This new algorithm helps the humanoid robots “Atlas” to get a better feeling of the ground. Atlas is able to dynamically adjust its balance, can run and may stand on one leg even if poked. However, stability is an issue on uneven surfaces as impressively shown at the  DARPA Robotics Challenge where robots have to prove their off road capabilities. Here tasks like climbing over rocks or getting out of a car have to be fulfilled.

The software of Atlas tries to foresee problems by carefully approaching difficult terrains. First the robot puts his leg on the ground and tests, if it is stable. Then he moves forward and – if necessary – adjusts torso and arms. In a paper the scientists describe that the robot tests stability by first examining the contact around the leg. Depending on the leg rotation stability is derived. This approach is purely software based, no additional hardware is needed. Currently it remains unclear if the technology will also be commercialized.

This entry was posted in Artificial intelligence, robots. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s