The MIT Professor Obsessed with Building Intelligent Prosthetics

One Saturday morning in January 1982, Hugh Herr and Jeff Batzer set out to climb Mount Washington, the highest peak in the Northeastern United States at 6,288 feet.

To even the most dedicated climbing aficionados, it appeared a foolhardy mission. Avalanche conditions had swept through every gully on the mountain except for one, but Herr, 17, was well known for his fearless spirit.

A climbing prodigy who scaled the face of Mount Temple, a 11,627- foot peak in the Canadian Rockies at the age of eight, Herr had even begun free climbing without ropes in his early teens. Such was his reputation that Batzer, despite being three years his senior, trusted him implicitly.

But 1,000 feet from the summit, things would start to go very wrong. With winds reaching almost 100 MPH and visibility no more than a few feet, the pair were forced to seek shelter on the Great Gulf, an exposed glacial cirque on the mountain’s north face. With no way up or down, they were trapped for three nights, exposed to the full force of the elements. By the time a search party found them, both were close to death.

Herr’s legs were severely frostbitten. After seven surgical attempts to repair the circulation, doctors informed him that they had no choice but to amputate both limbs below the knee.

“As a double amputee back then, you were regarded as almost worthless by society,” Herr said. “Someone spoke to my father shortly after my legs were amputated and asked, ‘Is Hugh married?’ And my father was slightly confused and replied, ‘No.’ And they replied, ‘Oh good, thank god,’ as if I’d become subhuman and how horrific it would be to be married to me as a cripple.”

Far from being crippled, Herr would soon climb again, faster, higher, and better than before.

Over the past two decades, as a researcher and professor at MIT, Herr became one of the world’s leading pioneers of bionic, “thinking” limbs—prosthetics embedded with machine intelligence that may become more agile, mobile and useful than their biological equivalent. Herr believes that they will eventually change the way the entire human race moves, travels and even looks.

Ten years ago, these ideas were the realm of sci-fi films. To most of us they still seem barely conceivable, more Iron Man than everyday life. But in a few select human test subjects around the world, they are rapidly becoming reality.

Artificial intelligence today falls far short of mirroring the complexities of the human brain which contains approximately 1 billion neurons, with around a trillion connections between them all. So far the closest we’ve come to replicating that in any shape or form is the OpenWorm project, in which researchers created a computer simulation of the 302 neurons found in a roundworm, and used this model worm brain to power a Lego robot.

Recreating human-like intelligence or even consciousness on a computer remains elusive, but what we do have available has begun to transform the lives of many amputees.

Forty-eight-year old Gummi Olafsson acts as a test pilot for Icelandic company Ossur, one of the world’s leaders in prosthetics research. At the age of just nine he was struck by an oil truck. After over 50 operations, he eventually opted for a lower leg amputation. “I had been in pain for 28 years,” he said.

Since his amputation, Olafsson has been wearing a bionic leg programmed with a kind of artificial intelligence that allows the limb to learn and adapt to his body.

Such bionic limbs and joints contain onboard computers which can sense the environment and flex the knee or adjust the ankle angle to match different terrain. They learn their wearer’s unique gait within 15 steps, adapting leg swing and speed to the individual’s natural walking pattern, and continue to adjust over time. The learning algorithms also automatically add power to the joints to make it easier when the wearer climbs stairs, and they can even sense when the wearer is becoming tired, adding additional strength and stability.

These bionic limbs have become so advanced that many believe they will start to become superior to their biological equivalents within the next five years.

To read the rest of this article, published in Motherboard, please click here.