MIT’s Cheetah 3 robot already climbs ladder without visual aid.

System used by the robot calculates probabilities with the help of gyros, accelerometers and positions of the leg joints.

As for us, the visual recognition of an environment is essential for the movement of robots, even the most sophisticated, such as those of Boston Dynamics. But what happens in case of sensor malfunction or lack of visibility in some place? It was thinking exactly in the absence of the “eyes” that the team responsible for the “metal labrador” or Cheetah 3, at the Massachusetts Institute of Technology (MIT), created a system called “blind locomotion.”

The Cheetah 3 weighs almost 88.185 Ibs hence the nickname “labrador” 🤔 , and can climb stairs and uneven terrain or quickly recover its balance without visual resources, thanks to two new algorithms: one for contact detection and another for predictive control model. The first helps the machine determine the best time for a leg to change its swing in the air for contact with the surface. So she can decide for the best action by stepping on a fragile or tough object and adjusting her balance.

The second helps the robot to determine the best moment to make the transition between the swing and the pitch, constantly calculating for each leg three probabilities: that of making contact with the ground, that of the force generated by touching the ground and being in the through the air , watch  the video below to get an idea of what we are talking about 👇 .

This is all measured on the basis of data from gyros, accelerometers, and positions of the leg joints, which record the angle and height of the limb relative to the surface. If, for example, there is a false step, your body suddenly tilts, changing its posture. This process happens 20 times per second on each leg.

It is very similar to what we do when entering a completely dark room, for example. “If humans close their eyes and take a step, we have a mental model where the soil can be and we can prepare for it. But we also rely on the touch of the floor. We are doing the same thing by combining various data sources to determine the transition time, “explains Sangbae Kim, one of the project leaders at MIT.

Well, you can not rule out that the “metal labrador” is going to be used for a day hunting us. Joking aside, the idea is to train you to perform very dangerous, dirty, difficult or inaccessible tasks for humans. Or even for trivial things that, while simple, can take up considerable time

MIT’s Cheetah with head and eyes.

“The Cheetah 3 is designed to perform versatile tasks such as inspecting power plants, which involves various terrain conditions, including ladders, curbs and obstacles in the ground. There are countless occasions when we want to send robots for simple, rather than human, tasks. ”

Researchers will now present the capabilities of the Cheetah 3 in October at the International Conference on Intelligent Robots in Madrid. In addition to “blind locomotion,” the team will demonstrate improved hardware, including expanded range of motion compared to its predecessor, which allows the automaton to stretch back and forth and rotate from side to side – resembling a feline about to attack.


Leave a Reply





King in Africa: Man lives in Germany and commands his country through Internet

Samsung elevates its Brand in techonology by inaugurating the largest mobile phones factory in the world