The words “robotic exoskeleton” probably bring to mind futuristic soldiers and sci-fi flicks like Aliens, Iron Man, or The Wrong Trousers. But despite military efforts to create such technology, it might show up somewhere less glamorous first: nursing homes. Researchers in Italy and Switzerland have developed a prototype device that can detect a slip in progress and help its wearer avoid falling. If perfected, a system like this could one day help millions of elderly people and amputees maintain balance and avoid serious tumbles.
As people age, they naturally become weaker and less agile. Add disease or injury, and falling becomes a worry with every step. And it can be more than an inconvenience—falls are the leading cause of fatal injuries among the elderly. Instead of waiting for the inevitable, some researchers are trying to take preventive action with powered exoskeletons—braces for the legs with motorized joints that assist while walking. But the braces are usually bulky and slow, and most people don’t need their assistance with every step. So the researchers set out to solve that problem with a device that would take action only when needed.
“It’s the first time that someone has rationally dealt with falls by having the robot collaborate with the person,” says David Reinkensmeyer, a biomechanical engineer at the University of California, Irvine, who was not involved in the research. “It’s supercool.”
The new Active Pelvis Orthosis (APO) consists of a waist brace holding motors on the hips that move lightweight carbon-fiber links connected to thigh braces. It uses an algorithm that monitors leg movement; if the legs diverge from a natural gait in a way that suggests a slip, the motors apply force to help the legs counteract the slippage.
To find out how it would work in people, its developers at the Sant'Anna School of Advanced Studies in Pisa, Italy, and the Swiss Federal Institute of Technology in Lausanne outfitted eight elderly adults and two above-the-knee amputees (who wore prosthetic legs) with the device. The device testers walked on a custom treadmill split in two. Once in a while, the right or left half would jolt forward, simulating a foot slipping on ice or a loose rug. Sometimes the APO was on, sometimes it was off, and sometimes the study participants didn’t wear the APO at all. Motion capture cameras recorded their limb positions, creating stick-figure animations for analysis.
After the start of a slip, the APO reacted within a third of a second, correcting a person’s gait for a quarter of a second. Stick figure analysis showed that—without the help of an additional restraining harness to prevent real falls—they would have fallen without help from the APO, the authors reveal today in Scientific Reports. What’s more, during normal walking, the APO, which weighs about 5 kilograms, had no effect on gait. “It’s a great example of trying a unique approach to exoskeleton control that’s having real results,” says Daniel Ferris, a biomedical engineer at the University of Michigan in Ann Arbor who was not involved in the research. “I’m very excited about it.”
The device doesn’t require much customization. After a user straps in, they program in their weight and take three steps. The device forms an internal model of their normal walking movement in just a couple of minutes. The device doesn’t completely replace a user’s reflexes, but merely amplifies their leg force by 20% or 30%. “It’s a nice example of a robot being synergistic with a person in an emergency situation,” Reinkensmeyer says.
Ferris explains why no one has done this before: There’s no universal hardware on which to test control systems, so each lab has to build its own. And you need experts in electronics, mechanical engineering, control algorithms, and biomechanics. In addition, he says, researchers have previously focused on the “low-hanging fruit” of helping users walk steadily, without worrying about recovery from falls.
Ferris thinks a device will be ready for the market in 10 years; Reinkensmeyer in just one or two. As part of the needed development to reach that goal, Silvestro Micera, one of the study authors, hopes to make the APO less bulky and untether it from the external computer that controls it. Future iterations will also include algorithms and motors to help with different types of falls, such as stumbling sideways, he says. If all goes well, your grandfather could become Iron Man before you do.