Podcast: Harvard's Soft Robot Gives Movement Back to ALS & Stroke Patients

In this episode, we cover Harvard’s soft wearable robot that learns a patient’s intent and provides just-enough support, helping stroke and ALS patients regain independence in everyday tasks.

This podcast is sponsored by Mouser Electronics. 

Episode Notes

(2:48) – Harvard’s soft robotic sling for stroke & ALS patients

This episode was brought to you by Mouser, our favorite place to get electronics parts for any project, whether it be a hobby at home or a prototype for work. Click HERE to learn more about how wearable robotics are helping Parkinson’s patients walk more freely!

Become a founding reader of our newsletter: http://read.thenextbyte.com/

Transcript

So, lots of folks with ALS or with strokes have a lot of difficulty moving their body. And the previous technology out there is really, really struggling to understand the intent of the user and provide the correct amount of support. But imagine if there was a robot that understood intuitively what your body was trying to do and then provided the correct amount of support. That's exactly what we're talking about today. A team from Harvard University built a wearable piece of technology that learns how each person moves and supports them with the correct amount of support that they need.

What's up friends, this is The Next Byte Podcast where one gentleman and one scholar explore the secret sauce behind cool tech and make it easy to understand.

Daniel: Hey friends, on today's podcast, we're talking all about wearable robotics to help folks with physical disabilities overcome a lot of the challenges that they have with moving their bodies, which is a really noble cause. And quick way to get an awesome primer on this type of technology comes courtesy of today's Mouser Electronics. We love working with Mouser because not only are they one of the world's largest electronics distributors, they're also our favorite electronic distributor for all sorts of projects we do. I actually about to order something from Mouser right now for work. But outside of that, they're also really well connected with technology. They know what's going on in the tech realm. They know new tech trends that are coming and they write awesome resources that help folks like you and I understand what's coming up on the horizon. So that's why we are happy to partner with them on a podcast and why we're excited to share one of their technical resources for today's topic, which is wearable robotics for Parkinson's. This article really goes over how soft robotic apparel can be used to help Parkinson's patients overcome gait freezing. This is specifically talking about how folks with Parkinson's have trouble with hip flexion. So, moving their leg, bending it at the hips, assisting people with Parkinson's that are having trouble with hip flexion, they give them sensors and cable-driven actuators as a part of the soft robotic apparel and it helps prevent their gait or their walking, their steps from freezing. Basically, completely eliminated it indoors helps reduce it outdoors. And it's an awesome understanding of kind of the principles of what we're talking about for today's podcast episode, which is using soft wearable robotics to help folks that are having challenges with moving their bodies due to physical disabilities to help them overcome those, assisting them in a smart way using robotics. So awesome technical primer for today's topic, awesome noble cause to be working on the robotic space and obviously awesome sponsor. So, we recommend checking out this article that we're linking to show notes today.

Farbod: It's also funny that the article linked in the ad is from Harvard and then we're also talking about Harvard today.

Daniel: Super locked in.

Farbod: Really focused.

Daniel: Unbelievably locked in right now. Let me mention though. So, you mentioned Harvard, today's topic is about using wearable soft robotics largely in the focus of shoulder and arm mobility. So, the previous topic we talked about was just hip mobility and leg mobility, shoulder and arm, different joint, similar principles. Lots of folks that are experiencing challenges with moving their arms, especially at the shoulder joint, they've had a stroke recently and they're having trouble regaining their movement after that or they've been diagnosed with ALS which is hurting their movement at the shoulder, which cascades down to your elbow and to your wrists. So, people who are having trouble lifting, holding, or lowering their arms, this team from Harvard built a soft robotic solution that they put on over their shoulder as a sling kind of support or brace almost. And it helps folks with strokes or ALS, move their arms, move their joints to help them to lift, hold, lower their arms. And it doesn't do the movement for them. It just supports them just enough that they're able to do the movement that they want to do when they want to do it. What's interesting is they also partnered with Harvard Medical School and Massachusetts General Hospital to get feedback from real physicians, but then also feedback from real patients who are experiencing these challenges and help them with the testing and development process, which I think is a great way to go about it as opposed to developing something independently in the lab and then hoping that it works when you get out into the real world. Use those connections, build some collaboration with folks who are helping treat these types of diseases in real time and patients who are volunteering and ready and willing to come test something new to help you develop something that really hits the mark.

Farbod: You just said something really important quickly where the device is providing just enough assistance to get something done, right? And that wasn't something they settled on right away. It's something they actually learned from previous attempts of creating these devices. They talked about how early on they had um devices that would predict motion. For example, like I want to move my arm up, right? And then the support from the robot would be like, great, get the arm up. And it would go like really up. And then the user might struggle to put it back down but the device doesn't really understand that. It's just providing maximum support. With this new iteration, they were able to take that feedback from the users and be like, OK, so we got to tone this down. We're going to do both physics, like a physics-based model and engine, to understand the minimum amount of support that we need to support a specific motion movement, and only provide that so that now we have a closed loopback system for the users moving up. And if they want to put it back down, where we understand the pressure that they're providing and how much we need to go down. I also wanted to point out that the article we mentioned from Mouser is Harvard Walsh Bio Design Lab. And what we're talking about today is also Walsh Bio Design Lab. So, we perfectly bridged ah the primer with the core article better than I think we've ever done before. Because now you clearly see how this design has been building on.

Daniel: This is going down in the books, for sure.

Farbod: It really is. It's an all-star episode already.

Daniel: I liked what you were saying though about the overreaction, the over-support. I can imagine it's challenging for me to understand, but when I kind of thought through analogy with myself, it's like, imagine if you're on the ground and my back's been bothering me a little bit recently, so there's been instances where I was sitting and I was like, my back really hurts. Nellie, could you just help me get up? Could you just help me get up a little bit? I'm little sore. And it's like, instead of her, gently assisting me to stand up, it's like, if she just yanked me as hard as she could and like, went flying, that's kind of what it feels like a lot of this technology has been in the past, which is like, Oh, like I want to lift my arm. And instead of just giving like just enough support to help me lift my arm on my own, the robotic system like yanks my arm up with full strength. That's not an optimal user experience and it doesn't actually capture the intent of what the user wants. So, I appreciate, like you're saying, their ability to get feedback from here.  One of the things that I want to highlight as well is the way that they ensured that they could capture the intent of the user correctly, which is called their intention detector model. And it's a machine learning model that they train uniquely for each user. They have to calibrate this model for each user. And it learns the user's movement patterns. It takes about 15 to 20 minutes. But within that 15 to 20 minute period, it gets up to 94% accuracy in trying to understand the user's intent on when they want to move, which type of movement they want to use, how fast they want to move, and how much they want to move, and how much support they'll need. And the second part of this comes from a physics model that's bolted onto the machine learning model, which they call their hysteresis model. And this helps it understand how much air pressure is needed to help support and lift and hold the arm depending on the movement that's intended. So, the robot combines both of these models and basically what it's doing is it's not only predicting what the user's intent is, it's also dynamically adjusting the amount of support that's provided to make sure that it's appropriate for that movement that the user wants. And after they do this 15-20minute calibration, it works really, really fast and really, really well. Again, with this 94-95% accuracy, it can detect the intent of the user's movement in under one quarter of a second and then reacts to providing full support in under half a second. So that's basically saying any time a user does any movement in less than one second, in about 0.75 seconds, you go from, hey, I'm wanting to do this movement, I'm thinking about this movement, I start to move, and in less than one second, the robotic system is already supporting you. And it uses less pressure than any other generation of this technology to help achieve these movements, which means that it's not replacing the movement of the user. It's just giving them just enough support to where they're able to do the movement on their own, which I think is really, really important, especially in the realm of strokes, where this can be used as a rehabilitation device. You don't want to be providing more support than the user needs. You actually want to be providing less and less support as the patient gets better and better and better. So, in this case is a really unique model where in addition to just understanding the intent of the user, they've also got this second physics model that's trying to minimize the amount of air pressure that's provided to help support the user to move. I think that's a really, really important distinction from a lot of the previous generations of this technology because the less force that's provided, the more the users are actually able to use their own body to do it. And even in the case of ALS, where it's a degenerative disease as opposed to something that's easier to be rehabilitated, the more the user can use their own body that can actually improve their quality of life and extend their quality of life when they're able to do the movement themselves as opposed to replacing their movements, their muscles with some outside force from a soft robotic support.

Farbod: Absolutely. Another thing that I think we just kind of quickly went over uh when we're talking about an upper body robotic mobility system. I think many people might be thinking of this hard rigid thing that has a bunch of motors everywhere, like an Iron Man version one suit. But this thing is actually just a normal vest, kind of, with pockets of air and a pneumatic system that can, like you were saying, provide the right amount of air pressure to float up and provide assistance to the joints that need it the most. I think that's actually a much better approach to design because like they said it in this article, patients that have ALS, so this is like a lifelong disease, or patients that are recovering from a stroke, the ability to have a supportive system that they can comfortably wear, like significantly impacts if they use it or not. And by looking at this design, you can clearly see that they've taken that into account. And this looks much more approachable than pretty much anything else I think I've seen so far. Another thing I was going to point out is, again, I'm a big visual learner. There is a video on this article where they show uh folks that have this system on with it enabled and with it disabled doing a task. So, I think one of them was someone combing their hair. The other one was someone trying to eat something. Just the amount of benefit you can visually see being added to someone's life is incredibly impressive.

Daniel: Yeah, for sure. And just like to hear actually to read, we'll link the article, which has the video in it, but read some of the quotes from, I think it was Kate Nycz was the patient who had been interviewed for the article. She's 39 years old, had been diagnosed with ALS at 33. And she was mentioning that it's affecting mostly her right arm. So, her right arm used to be her dominant hand or her dominant arm. And her arm, she said she can lift it to about 90 degrees from her shoulder and then it fatigues and falls. So, she's saying to do stuff like brushing her teeth, combing her hair where you're supposed to lift your arm above 90 degrees, it was basically impossible for her to do with her right arm. She's taught herself to become left-handed since. By using this device, she's able to now lift her arm to the angle she needs to be able to do it with her right hand as well. And this is six years into the progression of the disease. Imagine if she had something like this from the start. She probably could have improved her quality of life over the last six years to be able to use her dominant hand to do a lot of the activities she needs to do on a day to day basis. And it feels like small things, like brushing your teeth, brushing your hair, that feels like small things. But in terms of like your own confidence and your own quality of life and like your own dignity almost, like imagine not being able to comb your own hair or brush your own teeth or shave or having to teach yourself to do that with your non-dominant hand and then feeling confident to get up and go and like go out into public. That is a serious emotional and mental toll that's probably been grinding on a lot of these patients for years and years and years and years. So, it's like, again, it sounds trivial allowing people to comb their own hair, but the impact of that on each patient, on a day-to-day life, on a personal basis, I can only imagine the impact that this has on a lot of people. So, I'm really excited about seeing this technology get used on a sample size of more than just nine patients. And make it out into a world where it's a little bit more applicable to lots of patients with these types of these types of outlets.

Farbod: Couldn't agree with you more, man. Every time we do an article like this, I don't know, it feels really rewarding in the sense that it seems like as a community, the scientific world and the engineering world is moving in a direction that uh benefits those that need it the most. So, super exciting to cover something like this.

Daniel: Yeah, I'm with you man. Obviously, there's not always feel-good stories when it comes to technology and we need to measure our optimism on a lot of this stuff. But this is something that I'm definitely very, excited about.

Farbod: Me too.

Daniel: You cool if I wrap this up here?

Farbod: You want to wrap it up? Our minds think alike.

Daniel: Sweet. So, everyone imagine if there was a robot that knew exactly what your body wanted to do and helped you to do it before you even tried to move it yourself. That's what Harvard's new wearable technology is doing. It's helping people with strokes or ALS move more freely and overcome their difficulties with less effort. So, they have this soft shoulder robot that learns how each patient is moving using sensors and machine learning. And what's really unique about this one is that it adapts in real time. So, it figures out if the patient is trying to lift, hold or lower their arm and gives just the right amount of support, making movement easier and more natural for folks who are dealing with strokes or ALS.

Farbod: Love it.

Daniel: Thanks, Dawg. All right, that's the episode. See you around.

As always, you can find these and other interesting & impactful engineering articles on Wevolver.com.

To learn more about this show, please visit our shows page. By following the page, you will get automatic updates by email when a new show is published. Be sure to give us a follow and review on Apple podcasts, Spotify, and most of your favorite podcast platforms!

--

The Next Byte: We're two engineers on a mission to simplify complex science & technology, making it easy to understand. In each episode of our show, we dive into world-changing tech (such as AI, robotics, 3D printing, IoT, & much more), all while keeping it entertaining & engaging along the way.