Daniel Ferris says that the best exoskeletons are the ones you forget about. When the biomechanical engineer has one on, he notices little more than a slight tugging. But after removing the device, it can seem as if he is suddenly trudging uphill.

How a revolutionary technique got people with spinal-cord injuries back on their feet

Ferris has been designing wearable exoskeletons and powered prostheses for decades, and has long thought that they would soon become a part of everyday life. In 2009, he published a prediction¹ that “by 2024, people will be walking down the street, in the malls, and to their homes wearing robotic exoskeletons”.

Ferris admits he was “jumping the gun”, but his vision is getting closer to reality.

At the time he made his forecast, exoskeletons were mostly being used in clinical settings to rehabilitate people after spinal-cord damage or stroke. “Exoskeletons were really targeted as: function is lost, and we need to replace this function,” says Taylor Dick, a neuromuscular biomechanist at the University of Queensland in Brisbane, Australia.

Today, researchers and companies are increasingly talking about augmentation instead: adding a fraction of the force needed for a movement. A powered brace might take some load off an arthritic knee, and robotic shorts could make an older person’s daily walk less exhausting. Other exoskeleton devices are designed for a wide range of users, including young adults hoping to go farther and faster. Last October, Nike unveiled a prototype ‘powered footwear’ system called Project Amplify, which is aimed at everyday athletes, and tourists in China can rent exoskeletons during a visit to a hilly section of the Great Wall.

For proponents of exoskeletons, this augmentation can’t come quickly enough. Many countries, including Japan, China and Italy, are rapidly ageing, and the number of people worldwide aged 60 years or older is projected to reach 2.1 billion by 2050, according to the World Health Organization.

But before exoskeletons reach the mass market, researchers need to solve several technical issues. The devices need to become smarter, lighter, cheaper and easier to fit for the vast range of people who might use them. There’s also another concern that needs to be overcome: if a machine makes it easier to walk or run, will people move more and become healthier as a result — or will their bodies become reliant on it and their muscles atrophy as a result?

From replacement to assistance

For today’s exoskeleton builders, less is more in terms of the assistance the devices offer. Older devices trended more towards the style of Iron Man’s super-strength suit, whereas newer systems estimate the forces that the wearer is producing, and supply only part of what is required. That approach preserves the wearer’s control, reduces the power the machine must deliver and can make the device less taxing when it is switched off.

In the past few years, several teams have reported advances in developing devices that can support more-realistic movements. A control framework devised by researchers at the Georgia Institute of Technology in Atlanta uses an artificial-intelligence system to estimate how hips move in real time for a wide range of walking speeds, inclines and stair heights. The system enabled testers to reduce their effort while walking². Some of the same researchers and other colleagues extended the principle to a more flexible system that allows users to switch unpredictably³ between different types of movement.

Roboticist Robert Gregg at the University of Michigan in Ann Arbor is taking a similar approach with knees, and is leading a US$2-million project to adapt motorized orthotic braces to reduce the forces passing through the joint. In a four-person pilot study, participants reported reduced pain while using the devices during activities such as standing up from a chair and climbing stairs.

Robots rise to meet the challenge of caring for old people

The latest machines use what researchers describe as task-agnostic control. Whereas previous systems replayed a preset moving pattern that was timed to one body movement, task-agnostic control estimates the effort being made by the wearer’s joint in real time, then supplies a proportion of it.

Conventional exoskeletons have used small, fast-spinning motors and large gear trains, but the forces they produced could make joints feel rigid. Newer actuators use bigger, slower motors and lower gear ratios, reducing friction and allowing the wearer to move the machine more freely.

It all means that Gregg’s devices can provide around 25–30% of the force generated at a biological joint — enough, he says, to compensate for some deficits associated with ageing, osteoarthritis and repetitive work. Meanwhile, wearable sensors can track limb motion and how weight moves through a foot, then use machine-learning systems to infer the level of assistance needed. Gregory Sawicki, a biomechanist at the Georgia Institute of Technology, compares the goal with building “a large movement model” that is trained on enough examples of human motion to respond across users and tasks.

The result is less Iron Man than “wearable e-bike”, Sawicki says.

That analogy might sound odd, but it’s one that designers outside academia are echoing. Nike’s Project Amplify, developed with robotics company Dephy, in Boxborough, Massachusetts, uses a motor, drive belt and rechargeable battery worn around the lower leg and foot to assist ankle movement. It is still in testing, but Nike says that the firm eventually intends to sell it to people who want to walk or run farther and with less effort.

Alongside better actuators, soft materials are also moving exoskeletons closer to clothing. Researchers at the University of Heidelberg in Germany developed a system called WalkON that includes shoulder straps, a belt and bands over the user’s thighs — all of which fit over clothes. It reduced the metabolic cost of walking on level ground by 10% for ten older people. In a separate test, younger adults used 18% less energy while walking uphill⁴.

WalkON’s exoskeleton weighs just under 3 kilograms — much less than the cumbersome units of earlier generations of these devices. Dick recalls seeing a 23-kg unit designed for people recovering from a stroke. “Nobody can ever wear this,” she told its engineers.

Japan’s experience with robots and assistive technology shows why that matters. Tomohiro Shibata, a roboticist at the Kyushu Institute of Technology in Kitakyushu, tests robots that provide care for people, often in clinical settings. Cost remains the largest barrier, he says, but usability is also an important challenge. Some clinical exoskeleton systems take around 20 minutes to fit, he explains — which can waste time during rehabilitation appointments.

Like Sawicki, Shibata’s comparator for next-gen exoskeletons is an e-bike. An exoskeleton must become something a person controls “as if it is part of the body”, he says, without specialist help or constant maintenance. He says that the size and weight of exoskeletons can discourage their use, in the same way that stigma discourages some older people from using canes or hearing aids.