Live data from Hacker News

Researchers design wearable tech that can sense glucose levels more accurately

uwaterloo.ca

61–70 of 171 posts

Re: Researchers design wearable tech that can sense glucose levels more accurately

#61
post #41

Earlier quoted context omitted.

CGMs themselves are still inaccurate compared to needles. And even needle based meters have a lot of different levels of accuracy. A CGM, even once calibrated, may be off by as much as 10%.

The value is in clinical application though. For closed loop systems, making a clinical decision for insulin dosing every 5 minutes is life-changing, even if the dosage is 10% off. I’d say even with compression lows, most systems are self-balancing enough to produce better results than a self-monitoring person could. Besides, 10% off doesn’t often matter: At 0-70 mg/dl the pump should suspend insulin either way. At 1…

So maybe it gets paired with the self-attenuating insulin and the experience gets even better despite some inaccuracy!

Re: Researchers design wearable tech that can sense glucose levels more accurately

#62
What I miss in the paper is any accuracy figure for glucose sensing. If this is an alternative to needles, how would the measurements compare? That is the first question one should ask.

I fear we can assume that, although the approach might be novel, it can't replace needles for accurate measurement. But maybe I am overlooking the performance comparison.

Re: Researchers design wearable tech that can sense glucose levels more accurately

#63

How can an algorithm be "artificial intelligence algorithm"? Does it mean AI models are used for data fitting? Or clustering? For data generated in such low scales, wouldnt statistical methods or procedural methods be sufficient or efficient or both?

As an engineer working in the field who has designed both cloud algorithms and on-firmware algorithms, when marketing uses AI, it tends to just be training/data fitting. At best, the most complicated ones tend to be random forests and if any use neural networks, it’s usually just overkill. The answer to your last question is yes, especially when it’s from raw signals. Tbf, there are applications from devices that do…

>when marketing uses AI

In my experience when marketing wants to use AI, they will. Regardless of whether it is ML, basic statistics or even just a few if-else blocks.

It used to be the way you describe up to about 2-3 years ago, now the term is meaningless.

Re: Researchers design wearable tech that can sense glucose levels more accurately

#64

> Currently, diabetics must frequently prick their fingers or rely on invasive wearable patches with micro-needles to track their blood-sugar levels Type 1 diabetic here - for what it's worth, CGMs aren't particularly invasive. At least in comparison to the many many years of finger pricking! But a smart watch solution would be cool. (I actually do get my CGM readings on my smart watch, which is really nice!) I know…

Those might be really great for T2 though. I don’t need to know the exact number just a ballpark number to know how I’m doing. Having always had a bad relationship with food, I fall off way too easily without a CGM so for me, those things would be perfect.

But also in times where we have the Libre 3 which is so tiny that you legit don’t even notice it, a CGM on your wrist is not worth the loss of accuracy for T1 I guess (assuming your insurance pays for it).

Re: Researchers design wearable tech that can sense glucose levels more accurately

#65
This sounds great for the problem they are trying to solve but it also sounds like it could be useful for a lot of other interesting applications. I would assume you could put the same sensor on a plastic pipe and infer details about the liquid inside. If I was this team, I might look at how to commercialize this tech for industrial process control in parallel with the slow process of commercializing it for medical purposes.

"The system’s key components are a radar chip, which sends and receives signals through the body, an engineered “meta-surface”, which helps focus these signals for better accuracy, and microcontrollers, which process the radar signals using artificial intelligence algorithms. The algorithms improve the accuracy and reliability of the readings by learning from the data over time."

Re: Researchers design wearable tech that can sense glucose levels more accurately

#66
Looks like it's using mm-wave radar technology to detect changes of the dielectric properties of solutions with different levels of glucose.

There is also a paper from them doing it with the Google's Project Sali dev kit, the radar that was mostly demonstrated with gesture recognition, but looks like it's useful for other things too like this. They also show it can be also used to detect glucose levels in drinks too like Coke vs diet Coke.

Besides the current one posted, they have multiple other publications about the topic:

Using Project Sali back in 2018: https://scholar.google.ca/citations?view_op=view_citation&hl...

This one is also nice detailed one: https://www.mdpi.com/2072-4292/12/3/385

and looks like the current one is about improving this technology with enhanced signal-to-noise ratio.

Re: Researchers design wearable tech that can sense glucose levels more accurately

#67

> Currently, diabetics must frequently prick their fingers or rely on invasive wearable patches with micro-needles to track their blood-sugar levels Type 1 diabetic here - for what it's worth, CGMs aren't particularly invasive. At least in comparison to the many many years of finger pricking! But a smart watch solution would be cool. (I actually do get my CGM readings on my smart watch, which is really nice!) I know…

> but for me, the biggest quality of life improvements for me will be continued improvements with closed loop CGM + insulin pump systems

You might find this interesting: "A bi-hormonal fully closed loop system"

https://www.inredadiabetic.nl/en/home-english/

Re: Researchers design wearable tech that can sense glucose levels more accurately

#69
post #66

Looks like it's using mm-wave radar technology to detect changes of the dielectric properties of solutions with different levels of glucose. There is also a paper from them doing it with the Google's Project Sali dev kit, the radar that was mostly demonstrated with gesture recognition, but looks like it's useful for other things too like this. They also show it can be also used to detect glucose levels in drinks too…

Nit: Google's project was called Soli.
Post reply on HN