Transcript: Space4U podcast, Robert Gregg & Toby Elery

Written by: Space Foundation Editorial Team

Hello, this is Andrew de Naray with the Space Foundation, and you’re listening to the Space4U podcast. Space4U is designed to tell the stories of the people who make space exploration today more accessible to all. Our guests today are Robert Gregg and Toby Elery who are part of a team that has designed an approved robotic prosthetic produced using parts originally designed for use on the International Space Station.


Robert Gregg is an associate professor of electrical and computer engineering and robotics at University of Michigan Ann Arbor. He earned a Bachelor of Science in electrical engineering and computer sciences from the University of California, Berkeley, and then subsequently master’s and doctoral degrees in electrical and computer engineering from the University of Illinois at Urbana Champagne.


He joined the University of Michigan as an associate professor of electrical and computer engineering and the robotics Institute in fall 2019. Prior to that, he was an assistant professor in the departments of bioengineering and mechanical engineering at the University of Texas at Dallas with an adjunct appointment at the UT Southwestern medical center.


He was also previously a research scientist at the rehabilitation Institute of Chicago and a postdoctoral fellow at Northwestern University. Dr. Gregg currently directs the local motor control systems laboratory, which conducts research on the control mechanisms of bipedal locomotion with applications to wearable and autonomous robots.


Our other guests, Toby Elery is a mechanical engineer and researcher based in Dallas, Texas. He earned his PhD in mechanical engineering from the University of Texas at Dallas in the spring of 2020. Up until April of this year. And for nearly six years prior, he was a PhD graduate research assistant at the University of Texas at Dallas.


His areas of specialty include mechanical engineering, research and development engineering, robotics engineering system, design engineering, and electromechanical engineering. And when Toby was finishing his undergrad degree in mechanical engineering, his eyes were opened to the world of medical devices when he had the opportunity to visit a prosthesis clinic.


And this inspired him to apply his engineering background to improve people’s quality of life. It also led him to pursue his PhD in mechanical engineering and to conduct research in robotic prosthesis and orthoses. He’s worked on a host of robotics projects has served as a mentor for undergraduate projects and has also disseminated his research in the field through several publications and presentations.


Robert and Toby, thank you both for joining us today. My pleasure. And it’s great to be here. So let’s start with how the idea came about to produce better robotic prostheses. Yeah. So that’s been the goal of my research lab, uh, since its beginning, actually I was kind of the foundational, um, uh, vision for the lab and it’s driven by a need for better prosthetic legs because conventional prostheses, although they’re designed and manufactured very carefully to restore as much function as possible.


They’re designed to be mechanically passive, which means that they can flex. They can absorb energy. They can, um, sometimes slow you down, but they cannot actually help propel you or help you get up out of your chair, help you push forward while you’re walking. And so that necessitates, uh, new robotic devices that can do that.


So when you introduce motors and sensors, Um, you were able to, um, have the prosthetic limb perform in a similar manner as the biological limb would do. Um, before the amputation, that’s really where our laboratory has come into the picture and has been helping push forward this area of research. Um, there, there are certainly other groups that are working in this area as well.


And we just happened to have a new way of designing a robotic leg using specialized motors that as you mentioned, were originally designed for robotic arm on the on the International Space Station. So who said like, aha, this motor from the ISS is going to fix this problem. How was that connection made?


Well, I’ll let Toby answer that. But first say that our goal with this study mine was to make it as dynamic and, uh, free swinging as possible because one of the side effects of using motors is that you often have to use gearing. And when you use, um, a smaller motor, you typically have to use a lot of gearing.


And then that gearing actually adds friction and inertia that, um, makes the joint very stiff, very, very rigid. And so it doesn’t kind of move naturally like your biological joint would essentially, it’s only moving exactly how the robot is forcing itself to move. And so what I essentially worked with Toby on was we need to create a new type of design that uses a different type of motor to minimize the gearing, um, so that we can have both the power that we need, the mechanical power from the forces that the motor generates.


While also having the benefits of free-swinging motion, much like a conventional leg paths and Toby would be the right person to, to answer how he came across the two specific motors. Yeah, definitely. So we picked these, it was years ago when we were first starting kind of the concept design of this prosthesis.


And, um, I was pretty early in my grad school. And I was working with who’s now, Dr. Hanqi Zhu, who was a former PhD student of Bobby’s as well. And he sent me this motor, um, he said, y’all take a look at it. And we started looking at it for the application that Bobby was just discussing. We wanted this really compliant actuator.


So we wanted a really high torque motor, um, so that we could kind of reduce the, the gearing necessary. And, uh, he sent me the, the motor and we were looking at it and we, it was probably one of the highest had one of the highest torque density, which means it can produce a lot of force in a very small package, which was really useful for our application.


So we were able to get a whole lot of torque, a whole lot of force out of it. In a really small volume, which is critical when designing motors into prosthetic devices that are, that are worn on people. We can’t have these enormous devices. It’s just unreasonable to wear that on, you know, something that’s too big on a person.


So that’s kind of where we found these motors and that’s kind of why we started using them just, just for how impressive they were. Do you know how they were used aboard the ISS, those motors? I don’t know a lot of specifics, maybe Bobby would, but I know that they were used, it sounds like in the robotic arms, but I have heard that, you know, they’ve got all sorts of robotics up there to help the astronauts, you know, do a variety of experiments or just different tasks that they have to do.


You know, having a third arm without having to send another person to space is always useful. So I can give a little, a little bit more background. So the, the, uh, German space agency DLR as they’re called, they, they originally developed these motors and, um, I don’t know which specific robotic arm it was for, but it was for some arm, I guess they probably have um, more than one or that they have developed more than one.


And when they had finished designing this type of motor, they had a spinoff company called Robodrive to commercialize it and distribute it for other robotics applications. And so we that’s where we came in, in touch with, with this type of motor was through, through Robodrive, um, which at the time that we were designing our leg, they were, they’re still very new.


Um, and not really widely used in the field. And so it was kind of a happy, lucky coincidence that we came across them at the time. And we didn’t even know there was this connection to space at the time that we did, uh, come across them. It was only, um, later we’ve, we’ve found out more about the history of the motors.


That’s how it came across. Interesting. So is it the motor itself is so much quieter or is it the gearing or is it a combination of both. Yeah. It’s definitely a combination of both. Yeah. Because the audio of the leg is reduced so much mostly because of, yeah, I’d probably say a lot of the gearing, uh, but that is enabled through the, the motor that we use.


So when you have what’s typical is you have these transmissions that have really high ratios. And because of that, um, there’s also usually a lot of friction and damping in the system. A lot of parts that are meshing with each other. Whether it’s, you know, physical gears, uh, gear teeth, you know, meshing together, or if it’s a lot of little ball bearings that are rolling, you know, every time if you have more of those, there’s more friction that has to overcome.


And just more of those rolling, which create this loud sound. Right. And because with this motor, with a reduced transmission, we’re able to kind of minimize the amount of parts that were necessary. And the transmission, so that really helped us reduce the audio of the leg as well. It’s a bit more compact, more powerful motor allowed you to eliminate gears.


Yeah, essentially. Yep. All right. Um, yeah, we still, we still had a minimal set of gears. Um, essentially, what is it? Toby at 24 to 1 gear ratio. 22 to 2, uh, off by two. Um, yeah, so, um, there’s still some minimal gearing, but, but this that’s a order of magnitude smaller than the conventional designs, um, that we’ve seen in, in the field of, of robotic prosthetic legs, which typically use 200 to one or more.


Um, and they’re, they’re, um, they’re gearing, which, and w what’s interesting is that, um, That doesn’t mean that there’s 10 times more inertia. That actually means that there’s a hundred times more inertia because the re nurser reflected through the transmission scales with the square of the gear ratio. So, um, there’s actually a really substantial effect by minimizing the gearing.


It’s not, it’s not a proportional effect. It’s actually a squared effect. I’ve heard that, uh, conventional prosthetic wares have to exert about one and a half times the amount of energy to operate their prosthetic compared to utilizing a natural limb. Is that accurate? And how much does your prosthetic lessen that burden?


It can be even as high as two or three times more metabolic costs depending on the level of amputation. And so essentially the reason for this is that when, so someone who’s lost, their limb is walking with a conventional leg. Uh, as I discussed earlier, the conventional leg does not produce the function of the, of the missing muscles.


Um, it doesn’t help inject energy and help propel you. And so what ends up ends up happening is that you’re attached limbs like the hip on the amputated side or the sound leg, the attached leg ended up compensating to, um, insert that energy, um, that is missing from the prosthetic side. And that’s a, that’s not a natural thing that our, our biology is designed to do or, or evolve to do.


And so it’s less energetically efficient to, to do that. So essentially it’s less energetically efficient to, to propel your, your walking with your hips. Than it is with your ankle. And so that, that’s the reason that it’s more energetically costly. Um, now we have not actually tested metabolic consumption, so that, that requires specialized equipment using, uh, essentially measuring the CO2 that comes out of your lungs through your mouth to, uh, measure caloric consumption.


But, uh, so we haven’t done that testing yet, but what we have seen is that we can reduce the amount of mechanical work, essentially the amount of energy that the hips are injecting into the gait cycle compared to when using a conventional passive leg. So the power leg minimizes the need for those compensations.


It doesn’t completely eliminate them. But it’s minimizing that the use of that compensation, thanks to restoring more normative biomechanics in the, um, in the prosthetic leg using the powered joint. That’s great. And I can add just a little to that. Okay. Uh, yeah, like Bobby said that we didn’t measure the metabolic energy used, but we did measure the mechanical, which is based off of, uh, the joint torque and velocities.


And we were able to see roughly on average, about a 13% decrease of work at the residual hip. Uh, which is where a lot of the compensations that Bobby mentioned come into play. So, yeah, we, and that’s not necessarily 13% below what is normal and healthy individuals, less 13% below what they’re doing on their typical daily prosthesis.


So we’re, we’re going in the right direction. And hopefully one day we’ll get to where, you know, it’s a seamless integration and. You know, they’re, uh, not having a compensator, uh, expend any additional energy at all. Sounds significant. Um, there’s also a secondary benefit of this is that, um, then they’re not overusing those, um, residual joints.


So, uh, over time, if they’re overusing their hips, they can develop osteoarthritis. Same with the, um, the sound leg. Um, if they overused their knee joint, um, they can develop, uh, osteoarthritis there too. And so, um, actually it’s not just making, walking more energetically efficient for the individual. It’s also saving their joints so they don’t deteriorate over time.


Um, now again, this technology is too new to be able to say definitively that this will prevent that, um, at this point, it’s we have anecdotal evidence to suggest it. We just know that we’re going in the right direction based on the root cause of the overuse. And how extensive has the testing been thus far, are you finding it’s universally beneficial for like different kinds of wearers, or?


So our testing so far has been pretty limited. We’ve only tested on three, uh, amputee individuals. And so, like Bobby said, the, a lot of the stuff that we’re saying, can’t be really stated as fact across all the amputee community. But, and, and there are some things that we’re saying is, you know, more on the individual level on how it affects the person specifically.


But one of the things like we mentioned before, uh, we do see, like Bobby mentioned that these devices can inject energy to help push you forward. We’re seeing that which should happen. We’re seeing a reduction in hip work, kind of across the board. We’re seeing some increased cemeteries across the board, too.


So it definitely depends on the individual. Um, and we can’t make any huge statements on how it will affect the community as a whole. But again, I think we’re, we’re going in the right direction. How are they attached to the residual limb? Is it a harness or pin lock or suction? Yeah. So this is a knee ankle prosthetic leg that we’ve developed.


So this is for persons who have an amputation above the knee. Um, and what they typically have is a socket that fits onto their residual. And, um, and like you mentioned, it, it. Uh, stays attached through suction. Uh, so they have, these are typically made with their clinicians and, uh, that’s what they use with their daily prosthesis.


And then the way we’ve designed our leg is they can easily take off their daily prosthesis. And then. Uh, very easily put on ours. So it’s just a really quick, uh, you unscrew a couple bolts and then you Holton, uh, the new prosthesis being that it’s an ISS motor or produced for that. Will they be affordable and practical to produce, do you think with large enough scale?


Yes. And I think that we’re starting to see. Additional motors, additional companies producing similar style of motors that are working at increasing the torque density. And when they’re doing that and when they can do it at a large enough scale, I do think they can get it to where it’s a more cost effective to the individual.


The ones, when we purchased them, we only bought two motors ‘cause we needed one for each joint. So. At that scale, they were not cost-effective for, you know, mass distribution. Uh, they were between one and $2,000 each if I remember correctly. Um, but one of the nice things like Toby was getting at with the other companies that are trying to build and distribute similar style motors, that at a larger scale, um, that’s, that’s largely driven by the drone industry.


So there’s a lot of cool technology that’s becoming widely available, thanks to hobbyists and professionals that use that use drones because they also need to generate high torque. Uh, they also need to be very lightweight because the drones need to fly and they also need to be cheap because, you know, if the drones are being used for, for a hobby, you typically don’t want to throw 10 K into it.


Right? So we are as a field, we are greatly benefiting from. Um, better motors, better sensors, better computation that lighter, cheaper from parallel industries like drones, uh, electric cars and so on. That’s great. The another industry that I recently kind of looked into are electric bikes. Um, they are starting to implement what we call direct drive motors, which is just a motor without a transmission.


Which, you know, ours is kind of an in-between. We have reduced the need on the transmission, but we haven’t eliminated it completely. But with bikes, they have a little more room to work with, so they can kind of eliminate the transmission, but I’ve seen some electric bikes just have a really big direct drive motor on them.


So it’s somewhat different application, but similar that they are improving on motor technology as well. Yeah, the same needs, the lightweight and high torque. So another aspect of the design, as far as I understand, is that the force from the residual limb actually charges the battery. Uh, how does that work?


Does that, uh, take care of the charging process essentially? Uh, or does it require additional charging? Yeah. So I guess with your first question, how does it work? We’re talking about the initial kind of design and desire for the style of prosthesis, as we want it to minimize the, uh, the impedance or the reflected inertia of the joints.


And so essentially we want them to be very flexible, right? So kind of like, you know, our, our joints, um, if you were just to like lay limp, someone could come up and easily move, you know, your knee or your foot. Where that’s not the case with a lot of robotics, they’re usually very stiff. And so with ours, because we’ve made it more compliant, you could easily come up and you can move those joints.


Just like we’re very similar to, you know, an actual anatomical joint. So because of that, uh, any time the device impacts the ground, it moves the joint and kind of back drives the motor, essentially making it a generator. So, um, there, it can act in two ways or one where you send it electricity and it creates force.


And the other is when it impacts the ground, or maybe when you’re swinging it in the air, it can absorb the energy and acting as a generator. To go and recharge the batteries. Now that helps with things like efficiencies of the entire system, but the energy absorbed is not quite enough to fully recharge the batteries.


And so they are, they are rechargeable and they do need to be recharged. Okay. And, uh, you kind of touched on this earlier. I know that a concern for wearers of prosthetics is the size of the bulkiness of the weight. Uh, how does this compare to what’s currently available in those aspects? And like, for example, does it allow the wearer to wear pants?


Yeah, so, so I would say that this design is most definitely heavier than conventional prosthetic legs. And slightly bulkier, but not too much bulkier. So the weight of course comes from the fact that we have motors and batteries, right? Those are the heaviest components. Of course, all legs are going to have some sort of metal or carbon fiber for the structure.


Right. So that, that, that you can’t really get rid of. Um, but we, in addition to that, we have motor sensors, controllers and batteries. And so it does, it does, you know, increase the way. But it’s not, it’s not heavier than a biological limb would be, but that still doesn’t mean that we can declare victory though, because, uh, keep in mind that when you’re wearing a prosthetic leg, it’s not directly interfaced or attached to your, to your skeleton.


It’s actually hanging off of soft tissue on your residual limb. And that’s being held on by suction like, like Toby mentioned earlier. And so you don’t want to have a lot of weight hugging on, on your skin that can cause discomfort, sweating, pressure, sores, um, and it’s just doesn’t feel good. And so our goal is a quintessential engineering trade-off like you want to have high-performance.


But you want it to be as light as possible and you can’t get both. Right. But we found a kind of a nice balance compared to other robotic legs that use different styles of motors. Some robotic legs are lighter than ours, but then don’t produce as much force as ours can.


And they’re also not as free swinging, like we mentioned earlier. And so again, these are all just trade-offs. You have to consider imbalance and what we’ve been arguing and hopefully showing initially through our results is that the extra power and propulsion that the leg gives you. And the fact that it’s free swinging, is it actually enough to make up for the extra weight, especially while walking, if you’re just standing still and trying to lift up your leg, then you’re going to feel that weight, right?


There’s no, no way around it. But if you’re walking and the leg is actually helping push you forward and help push itself off the ground into the swing phase, then you’re not actually going to be feeling that weight as much because the leg is actually helping you do that. And your hips can just kind of go with the motions, go with the flow and much like when you’re walking with your biological limb.


So, um, yeah, in summary it’s, it’s still a challenge. We’re still trying to reduce the weight as much as we can, and we’ve made some progress with that. But, um, we do believe that that this approach is going to be effective in the long run. Um, as we continue to improve on, on our designs, on the components and make them lighter.


So the, the performance and the like momentum offset that weight. Yeah. That is definitely the, the idea. And we have, again, some preliminary evidence to suggest that that’s great. So with the amount of electronics going on, is there, is there any water resistance? Not right now. No, we stay away from water.


Okay. Is this prosthesis recommended for all types of wearers and activity levels? I know there can be some differences with that as far as higher activity levels versus mild activity levels. Okay. So this is, um, yeah, it’s a great question. Typically, your activity level is what determines what an insurance company is willing to pay for in terms of the type of prosthesis.


And so the more intelligent prostheses that are available on the market that have like a variable damping, which is really helpful for walking at variable speeds and having more, more natural emotions, those are typically reserved for the high-end activity individuals. So, um, there’s different, uh, levels called K levels, you know, 1 through 4.


And so a K3 is someone who walks in the community. Um, in addition to in home and is able to freely navigate the environment, they’re the kind of person that an insurance company would say. Yeah, you’re going to actually use this, this expensive device to improve your life. Um, so they’ll pay for it.


The K4 is like an athlete, you know, a paralympian kind of person. Right. Um, there are very few, um, uh, who, who meet that classification and then K1 would be someone who wouldn’t get a lot of use out of the prosthesis. And so. Cause they’re they’re limited mobility. And so, uh, typically insurance company won’t pay for a high-end prosthesis, like what we’re designing, but there, there is something to be said though, about technology being able to improve someone’s activity level.


So that’s something that the field is beginning to grapple with right now is, is as if you were able to provide better technology for someone who is a limited mobility individual. Would it actually help them improve their K level. And there’s some evidence of this with commercially available microprocessor knees.


So again, these are the variable damping designs. They cannot inject energy like our robotic design does, but they can, they can do nifty things like change their dancing between stance and swing. So there’s been some, some research by the Rand group. That demonstrates that actually, um, users can be improved from K2 to K3 using these emerging technologies.


And so, um, it’s yeah, we’re kind of in a neat debate right now about, about who is the appropriate end user for these devices. But I, I would say that our devices it’s, you know, a little bit bigger and a little bit heavier. You’re not the right device for, for someone who is elderly really weak and not able to, to, you know, to effectively wear it.


You know, that person would probably need something much lighter, much less powerful. Right. And like how about a higher weight group? Just kind of a similar thing, right? Yeah. I mean, our, our device would be very useful for someone who, um, who, you know, needs more forced to, uh, more, more assistance to get up out of a chair to stand up and to go upstairs and so forth.


Um, that would definitely help them. But, um, also keep in mind that some of these designs can be scaled down to a degree. You know, we could, we could potentially select a smaller motor. And smaller structure and so on. Um, we have not looked into that yet, but you could potentially imagine having like a, you know, a lightweight version and medium sized version and a heavy, heavy version for depending on people’s needs.


Speaking of different versions, you know, would for below the knee amputees would a foot with an ankle motor design be a possibility with this technology or this design. Yeah, definitely the way we’ve designed this prosthesis, it’s kind of modular. Such that the MI is kind of its own actuator thing and in the ankle and foot is kind of its own separate thing.


So future designs could take the knee actuator out. It would need a little bit of redesign just for electronics and packaging, but, um, yeah, it’s very possible. Great. I do want to point out that there is a commercially available, uh, robotic ankle prosthesis. Um, it came on the market a few years ago.


Hasn’t been super widely adopted yet. I think there’s a, you know, there’s a lot of, of, uh, obstacles to overcome, like, you know, insurance isn’t willing to pay for all of it. It’s, it’s, you know, it’s more expensive and the outcomes haven’t been well established yet, um, and, and studies of, of, uh, clinical outcomes.


And so, again, it’s very recent technology. It’s still emerging. And also just keep in mind that not all robots are the same, um, you know, And we talked about some robots that are highly geared or more stiff. Ours is more compliant. And so our ankle prosthesis, um, the, the module, uh, would be quite different than the one that’s available on the market.


And so it could have different outcomes. Um, but we have not yet investigated that. Cool. So I just have one final question for you guys. You now have a patent pending on this design. What’s the current outlook for manufacture? Are you currently shopping for a producer? Yeah. So the University of Michigan and the University of Texas at Dallas are working together to try to find someone to license the technology.


If that doesn’t happen, then there’s also a possibility of spinning off a startup company. Of course, that has its own challenges associated with it. Um, and it’s not something that’s, uh, at least I can do personally, given that I’ve already got a, you know, more than a full-time job as a lab director, but yeah, so essentially, we’re trying to find part commercial partners who would.


Who would bring it to market with studio, some sort of licensing agreement. Well, um, you know, that was all I had. Uh, it, was there anything else that, you know, podium’s yours, if there was anything else that I left out that you I’d like to cover? Well, I think, uh, um, just want to point out for any, any of our younger listeners that, uh, this type of work is interdisciplinary and there’s not one perfect path to working on robots.


So, uh, I’d say anything in engineering or computer science is certainly useful. Um, but, uh, you know, we have, we have lots of different skill sets in our group. Um, we have electrical engineers, mechanical engineers, biomedical engineers, robotics. Some universities have actual robotics programs like Michigan, uh, and we all work together.


And so, you know, being really good at your, what interests you and then working in teams is what it’s all about. Awesome. Well, uh, Bobby and Toby thank you for your time today. We appreciate you both filling us in on your awesome product, and we wish you the best of luck on its future production and success.


All right. Thank you so much. Yeah. Thanks. It was really great to join you today. And, uh, that concludes this episode of the Space Foundation’s Space4U podcast. You can subscribe to this podcast and leave us a review on Podbean, Apple Podcasts, and on Google Play. Don’t forget to follow us on Facebook, Twitter, Instagram, and LinkedIn.


And of course our website, where you can also learn about the various ways you can support the Space Foundation. And all of these outlets and more, it’s our goal to inspire, educate, connect, and advocate for the space community. Because at the Space Foundation, we will always have space for you.


Thanks for listening.

Listen to the Podcast

Space4U Podcast: Robert Gregg & Toby Elery — Robotic Prosthesis Built with ISS Motors