Listen to this text
A smoother expertise for robotic prosthetic leg customers is the purpose of a University of Michigan venture that has acquired renewed help from the National Institutes of Health. The R01 grant of $3 million will even allow the implementation of the researchers’ improved management program on a commercially accessible robotic prosthetic leg.
Sit to face, stand to stroll, up and down stairs and inclines — the human physique shifts amongst these actions nearly thoughtlessly. But these transitions are onerous for robots, and robotic prosthetic legs have the extra problem of not being hooked into the human consumer’s central nervous system to remain synchronized with their physique.
Working on the issue since 2013, principal investigator Robert Gregg, an affiliate professor of robotics, noticed early success controlling the place of the knee and ankle joints via a mannequin that repeatedly represents all phases of the gait cycle. Previously, robotic prosthetic legs used separate controllers for every stage within the gait cycle just like the heel strike, push-off and swing. As a end result, the management parameters for every mannequin and guidelines for switching from one mannequin to a different needed to be optimized for every affected person.
“Every individual has totally different parameters as a result of each individual walks otherwise. And that resulted in very, very cumbersome medical deployment,” Gregg mentioned.
Using the movement of the thigh to foretell joint place with a steady mannequin turned out to be a reasonably great way of making a pure gait. With the preliminary R01 grant in 2018, Gregg’s workforce prolonged the strolling management mannequin to cowl different necessary situations: inclines, stairs, sit-to-stand, and stand-to-walk. Study contributors might do all of those actions with extra typical biomechanics utilizing the robotic leg than they might with passive prosthetics. However, pinning the management algorithm to joint angles results in a extra inflexible expertise when attempting to vary actions.
“The robotic has very sturdy motors, and so, if you happen to’re controlling the place and it’s one way or the other incompatible with the surroundings, it will probably really feel very inflexible and jarring,” he mentioned, which may imply ache for amputees the place the limb meets the socket.
So Gregg’s workforce is now taking a look at controlling joint place not directly — mimicking biomechanical impedance as an alternative — utilizing their steady modeling framework. With an impedance strategy, there’s an equilibrium place, and the forces are set to softly pull the joint again into that place if it’s disturbed. This gives some wiggle room.
Gregg compares it to the suspension on a automobile. “You have the spring, after which you might have a shock-absorbing mechanism. You need to hit the pothole and have sufficient bounce to melt the jolt, however you don’t need to oscillate eternally.”
This ought to allow the leg to supply the identical skill to seamlessly transfer from one exercise to a different whereas additionally offering a cushier trip.
The programming of the leg relies on biomechanical measurements of individuals with two organic legs, to copy the movement that the hips and again advanced for. Users of passive prosthetic legs typically expertise ache within the hips, again, and natural knee as a result of method they should compensate for the lifeless weight of the prosthetic leg.
To find out how the leg ought to behave throughout actions, Gregg united with collaborator Elliott Rouse, who research the mechanical properties of wholesome human gait.
“We receive the measurements for figuring out the biomechanical properties of the leg utilizing an exoskeleton,” mentioned Rouse, affiliate professor of robotics and mechanical engineering, and co-investigator on the venture. “The exoskeleton principally gives no help however often applies a fast perturbation that displaces the limb. From these measurements, we are able to decide the mechanical impedance, together with properties like stiffness, viscosity, and inertia.”
The management packages will first be examined on the robotic leg that Gregg’s workforce constructed in-house, with motors to energy each the ankle and knee. Then, to see if the brand new algorithms are prepared to begin serving to folks now, the workforce will check them on Ossür’s Power Knee prosthetic leg. In addition to measuring the biomechanics of research contributors as they stroll with the prosthetic legs, the workforce will gather formal suggestions to quantify will increase in consolation and reductions in ache.
The Power Knee has a springy, passive ankle with a smaller vary of movement, however it’s also lighter than the lab leg—and most significantly, it’s serving to folks stroll now. Gregg’s workforce is assured that they’ll modify their management mannequin to work on it, maybe bettering on the fashions Ossür is at present operating.
Editor’s Note: This article was reprinted from the University of Michigan.