Hugh Herr has spent the past decade building better prosthetic limbs, but this week was the first time he was able to demonstrate one of his own devices. A double amputee since he was in a rock-climbing accident in 1982, Herr unveiled his latest design at an MIT conference on Wednesday: a novel prosthesis modeled on the human ankle.

“This is the first prosthesis that allows for a humanlike gait,” said Herr, director of the biomechatronics group at the MIT Media Lab, as he walked around the stage during his presentation. He wore the new device on his right leg and his standard prosthesis, encased in a worn black shoe, on his left. “It’s strong enough to push my body forward and to propel me up stairs.” (A webcast of the entire conference, including Herr’s presentation, is available here.)

Most prostheses are passive, meaning that they lack the ability to actively generate propulsion during walking. Because of this, amputees typically expend 30 percent more energy on walking than do able-bodied people. And they often complain about lack of endurance, says Hilmar Janusson, vice president of research and development at Reykjavik, Iceland’s Ossur, which manufactures and develops prostheses.

For their new design, developed over the past two years, Herr and his colleagues took inspiration from how the foot and ankle work. When we walk, ligaments and tendons store energy that is produced when the foot hits the ground. That energy is then used to propel the foot forward. “The architecture of the human leg is gorgeous,” Herr said on Wednesday. “When you walk, there is an elegant transfer of energy from tendon to tendon.”

The researchers mimicked this strategy with a series of springs and a small, battery-powered motor. The kinetic energy of the forward motion of the walker is stored in a power-assisted spring that is then released to help propel the foot forward as it pushes off the ground.

The new ankle is about 20 percent more efficient than previous prostheses–a significant improvement compared with the 3 to 4 percent boosts the researchers have achieved in the past. “The difference is like getting on a moving walkway at the airport,” Herr said on Wednesday.

The team is now trying to make the device lighter and more robust. The researchers aim to have a commercial version available by early next year. “When this is optimized, we think amputees will be able to walk more efficiently than you can,” Herr said.

A similar device is being developed by Thomas Sugar, the principal researcher at Human Machine Integration Lab, at Arizona State University, in Mesa and colleagues, which will be used to aid both stroke patients and amputees. One version is a supporting device worn over the leg that is used for corrective or rehabilitation purposes.

Tests of the device on able-bodied people, carried out in partnership with the Walter Reed Army Medical Center as part of the Military Amputee Research Program, show that it can indeed match the power required for efficient walking. (Click here for video.)