Skip to main content

MIT’s Cheetah ‘bot walks up debris-littered stairs without visual sensors

After making its stage debut at TC Sessions Robotics in Boston last year, the third iteration of MIT’s Cheetah robot is back with some impressive new tricks. Associate Professor Sangbae Kim will officially demonstrate the Cheetah 3’s new capabilities at Madrid’s International Conference on Intelligent Robots in October, but in the meantime, we’ve got a […]

After making its stage debut at TC Sessions Robotics in Boston last year, the third iteration of MIT’s Cheetah robot is back with some impressive new tricks. Associate Professor Sangbae Kim will officially demonstrate the Cheetah 3’s new capabilities at Madrid’s International Conference on Intelligent Robots in October, but in the meantime, we’ve got a sneak peek via laboratory video.

The “Blind Climb on Stairs” portion starts around 1:48. It’s not exactly graceful, but it’s still probably a lot better than most of could do attempting to walk up a flight a with our eyes closed. Complicating matters are the small pieces of wood littering the steps, approximating some of the non-ideal circumstances the robot will have to grapple with during the search and rescue missions for which it’s designed.

The Cheetah is doing all of this without any cameras or other visual on-board sensors, using what the team refers to as “blind locomotion,” essentially feeling its way up the stairs. So, why rob such a sophisticated robot of something as seemingly essential as computer vision?

“There are many unexpected behaviors the robot should be able to handle without relying too much on vision,” Kim says in a release tied to the announcement. “Vision can be noisy, slightly inaccurate, and sometimes not available, and if you rely too much on vision, your robot has to be very accurate in position and eventually will be slow. So we want the robot to rely more on tactile information. That way, it can handle unexpected obstacles while moving fast.”

The robot utilizes a pair of new algorithms — contact detection and model-predictive control — which help it recover its balance in the case of slippage. The ability helps the robot determine whether to have a leg in the in air or on the ground at a given time, allowing it to tenuously but tenaciously ascend the stairs.

The ability, along with already showcased skills like leaping over objects and running up to 14 miles an hour, are all in service of Cheetah’s larger vision of aiding search and rescue missions. The robot is designed to enter areas that might otherwise be too dangerous for its human counterparts.

Data & News supplied by www.cloudquote.io
Stock quotes supplied by Barchart
Quotes delayed at least 20 minutes.
By accessing this page, you agree to the following
Privacy Policy and Terms and Conditions.