Ballerina Robot Full Of Stunts and Balance

//

When we think of artificially intelligent robots and machines, we often regard their algorithms as more intellectual properties rather than tools for physical applications. We know future AI bots will have the brains, but just how coordinated and graceful with they be? If the Cubli device is any indication, the answer is clear: AI machines of tomorrow will be as graceful as ballerinas.

PHOTOS: Tasty Tech Eye Candy Of The Week

Developed by researchers Gajamohan Mohanarajah and Raffaello D’Andrea at the Institute for Dynamic Systems and Control in Switzerland, Cubli is a small 6-inch cube equipped with three reaction wheels that spin and stop to control the movement and momentum of the block. Via a series of intelligent balancing algorithms and sensors, Cubli can effectively stand on its face, hop onto its edge and balance on one corner as if it were pirouetting. By coordinating all three movements, it can essentially walk. Cubli can even keep its balance on shifting inclines.

Each reaction wheel is attached to a DC motor and motor controllers, housed next to inertial sensors, a processor and batteries.

“Inertial sensor measurements are fused to get an estimate of the tilt in angular velocity of the Cubli. The hull sensors provide the angular velocity of the reaction wheels,” the researchers’ YouTube video explains. “Based on these estimates, a nonlinear controller is used to balance the Cubli on its corner, while the angular velocity of the reaction wheels is kept close to zero.”

BLOG: Sound Waves Levitate, Move Objects In Three Directions

While Cubli may look like a cute toy capable of doing a few stupid pet tricks, its technology is anything but child’s play. In fact, the diminutive cube represents a new breed of machines that will be able to actively react to its environment and adapt to its surroundings. If you’re interested in what this coming machine revolution will look like, D’Andrea can offer you a glimpse in the presentation below.

via Fast Company

Credit: Gajamohan Mohanarajah, ETH Zurich

DISCOVERYnewsletter
 
Invalid Email