Robotic Ankles Stabilize Amputees

Andrew Corselli

For the study, the researchers worked with five people who had amputations below the knee on one leg. Study participants were fitted with a prototype robotic prosthetic ankle that responds to electromyographic (EMG) signals that are picked up by sensors on the leg.

Study participants were then tasked with responding to an “expected perturbation,” meaning they had to respond to something that might throw off their balance. To replicate the conditions precisely over the course of the study, the researchers developed a mechanical system designed to challenge the stability of participants.

Study participants were asked to respond to the expected perturbation under two conditions: using the prosthetic devices they normally used; and using the robotic prosthetic prototype.

“We found that study participants were significantly more stable when using the robotic prototype,” said Co-Author Aaron Fleming. “They were less likely to stumble or fall.”

Here is an exclusive Tech Briefs interview, edited for length and clarity, with Corresponding Author Helen Huang.

Tech Briefs: What was the catalyst for your work?

Huang: Our motivation came from the current prosthetics. Most of the prosthetics — actually all the prosthetics, including those powered by a robotic device — on the market are just not able to assist that standing balance. But it's an essential factor; we call it the building block for every activity in daily life. You’re shuffling about or holding a kid — those require very good balance control.

Since the devices can't provide assistance on the ankle, the people who have limb loss use the intact side for a compensation strategy. So, then we said, “OK, well, how can we really leverage these powered, robotic devices to assist the person's standing balance.” We said, “Well, we have a better solution. Connect to the human nervous system, which controls muscle activity.” A neural signal is sent to a muscle for control. When we control our elbow, we will activate the biceps so that our elbow moves. It's the same thing.

Tech Briefs: What was the biggest technical challenge you faced throughout your postural control work?

Huang: The technology side is very simple to me, as an engineer, compared to a lot of the really fancy, AI-based learning autonomy. Neural control is very straightforward, so you connect with the human muscle activity to drive the motors that act on the ankle. That's it. So, the main challenge for me was we don't know if a person who lost a limb many years ago still has a capability to control their residual muscle.

So, basically, the human brain is the controller. After not using this muscle for many years, is there any biomechanical meaning remaining — can the brain still generate the activation signal?

Tech Briefs: Can you explain in simple terms how it works?

Huang: We have the robot, that is called a pneumatic muscle or artificial muscle. When the muscle is activated by the nerve signal, it contracts, generating a force to pull the joint to move. Our design exactly mimics the biological system. The artificial muscle, when it's activated, will contract, generate the force, and then pull the ankle up. How you activate that muscle is through the muscle signal.

The nerve says, “OK, I want to activate the toe up muscle.” The muscle will generate the toe up, use it as a muscle activity to drive the muscle around the toe, around the ankle, and then it will generate a force.

Tech Briefs: You’re quoted as saying, “We're now conducting a larger trial with more people to both demonstrate the effects of the technology and identify which individuals may benefit most.” How is that coming along? Do you have any updates you can share?

Huang: Yes, we are currently testing — recruiting and testing patients.

We have demonstrated the mechanics. In the paper, we talk about posture control and neuromechanics. But in the new study we will also check the person's cognitive function. So, to study, for example, how much effort they actually need, cognitively, to generate such processes — and other things like embodiment, do they feel the device become their own body, which is also another cognitive aspect for the user.

Tech Briefs: Aside from the larger trial, what are your next steps? Do you have any other set plans for research?

Huang: One of the challenges for our current setup is it's great in the clinic with lab testing, but it's not portable. We really want to push this as a commercial, portable device that people can use. My strategy is I’m going to design my own device, the electric motor.

The other way is maybe we can collaborate with a company that already has a powered ankle. That could require a lot of discussions with the company, but that's my essential goal.

Also, one of my current research focuses on closed loop One thing that is missing is the human cannot feel the ankle. They don't know what the device is actually doing, even though they're controlling it. That would require additional feedback from the foot. That feedback may potentially help them to better control their ankle.

Tech Briefs: Do you have any advice for engineers aiming to bring their ideas to fruition?

Huang: One thing that I wanted to mention for this project is we actually started from a hazardous concept to publish this whole paper. It took us maybe five years, so we actually piloted a lot, step by step.

There are a lot of details, and also uncertainty. There's a lot of risk; you don't even know if it will work or not. But if you stick with your idea and the continual piloting to find the best solution, you will get somewhere. This is something that I'm hoping to tell engineers. Don’t rush to get everything done and then say, “Look how it works.”

Be careful about it, but also don't give up — there are a lot of hiccups while doing research.

---