Skip to main content

Galago-Inspired Robot Sets Leaping Record

No video provider was found to handle the given URL. See the documentation for more information.

Ushering in a new era of agile robots, scientists have designed a unique robot that is the highest continuous jumper to date, surpassing the abilities of its predecessors to reach substantial heights after quick successive leaps. Their study, presented in the inaugural issue of Science Robotics, has created a new platform for exploring unprecedented styles of robotic locomotion.

Duncan Haldane, a Ph.D. candidate in the department of mechanical engineering at University of California, Berkeley and lead author of the paper, was inspired for the study when touring an urban search and rescue training site, where he noticed the rugged and obstacle-ridden terrain resulting from natural disasters like earthquakes. “In order for a search and rescue robot to move quickly across the many kinds of rubble, it has to be able to jump. And jump more adeptly than prior robotic art has achieved,” Haldane observed during a 4 December teleconference with reporters.

With this idea in mind, Haldane and his team sought to construct an agile robot that could accomplish multiple leaps more quickly and efficiently than previous robots in the field. But where does one start for such an ambitious creation? Robotic design often takes cues from living organisms, which are not only useful for confirming that a certain movement is physically possible, but also for drawing out primary blueprints of the robot.

"It is our intention that Science Robotics will bear the quality hallmark of the Science family of journals and cover both the traditional disciplines of robotics and emerging trends, such as advanced materials and bioinspired designs. It will also cover all scales, from very large systems to micro- and nanorobots."

Guang-Zhong Yang, Editor of Science Robotics, and colleagues in an editorial for the journal's inaugural edition.

Haldane and colleagues looked to the northern lesser galago, a small primate that attains incredible heights by sequentially bouncing off branches. Previous studies have shown that the galago can jump up to 1.74 meters (approximately 5 feet and 9 inches), accumulating a net height of 8.5 meters after five successive jumps. What’s more, it takes the animal only 0.77 seconds to perform these jumps.

One of the galago’s kinetic features that permits such unsurpassable agility is called “power modulation,” whereby the primate can jump with 15 times more power than what its own muscles can produce. This phenomenon requires a sophisticated system of storing, releasing and recovering energy quickly from the first jump to the last – a system that been difficult to quantify and carry out through robotics.

Haldane and colleagues tackled the challenge by first developing a way to measure power modulation.

“We needed a way to score how ‘well’ any animal can jump, and by comparing those scores, we can improve our design of jumping robots,” said Haldane.

His team formulated a metric called “vertical jumping agility,” or the ratio of the maximum jump height to the time it takes to complete one jump. “To have a high vertical jumping agility you have to jump high and do it quickly,” noted Haldane.

The galago has a maximum vertical jumping agility of 2.2 meters per second. “It’s fair to say that animals can outclass any robot when it comes to jumping,” remarked Haldane. Indeed, no contemporary robots can rival the small creature’s profound jumping agility.

Haldane and colleagues’ robot is no exception. However, the lightweight and galago-sized robot (appropriately named “Salto” from the Latin word for “I jump”), broke precedent by reaching 75% of a galago’s vertical jumping agility – the highest of any robot thus far.

The authors attributed Salto’s bounding success to a novel combination of fundamental force-driving mechanisms – most importantly, its specialized leg, comprised of carefully linked carbon fiber bars and aluminum pins, and controlled by a motor-fueled latex spring.

The leg was uniquely shaped by “mechanical advantage,” whereby torque, or twisting force, from the stretched spring was converted into a greater force on the leg to propel Salto upward “Just as a crowbar provides you with a lever arm to generate a large prying force from an initial smaller force, our linkage system is a force multiplier,” explained co-author Mark Plecnik, a postdoctoral scholar at UC Berkeley.

By incorporating mechanical advantage into Salto’s design, the authors effectively simulated galago-like power modulation. At the beginning of a jump, Salto’s leg stored the motor’s energy in a spring, and during subsequent motion, or the wall jump, the spring discharged that energy at a much greater power than the motor alone could produce – 2.94 times greater, to be exact.

Because of mechanical advantage, “Salto can jump up to a height of one meter in 0.58 seconds and be immediately ready to jump again without requiring a long recharge time in between,” said Justin K. Yim, a study co-author and Ph.D. student at UC Berkeley.

After measuring Salto’s performance over 10 trials, Haldane’s team revealed that Salto was able to achieve a vertical jumping agility of 1.75 meters per second, similar to that of the bullfrog.

“So what is it about Salto’s leg linkage that allows it to possess these wonderful features?  It is all in the geometry -- it is all encoded in the link dimensions.  And how do we figure out these dimensions?  Link dimensions are nothing but lists of numbers,” said Plecnik, thanking supercomputers and algorithms for sorting out the equations behind Salto’s assembly.

“This robot can dynamically use its environment to get places it wouldn't otherwise be able to reach,” said Haldane. This bio-mimicking robotic strategy, the authors noted, could help expand movement possibilities in other robots.