Researchers from the University of Waterloo in Canada have developed a new system for diabetics to monitor their blood sugar without having to prick their finger several times a day.

Using artificial intelligence (AI) and a radar device jointly developed by Google and Infineon, the researchers can detect changes in glucose levels without physically drawing blood.

“We want to sense blood inside the body without actually having to sample any fluid,” George Shaker, an assistant professor of electrical and computer engineering at Waterloo and lead researcher of the study, said in a statement. “Our hope is this can be realized as a smartwatch to monitor glucose continuously.”

The research is published in the International Journal of Mobile Human-Computer Interaction.

The study

The research began as a collaboration with Google and Infineon, which co-developed the radar device called Project Soli and sought input from select teams around the world on various applications.

The Soli system is a 60 GHz mm-wave radar that is intended for a small, mobile and wearable gesture-recognition platform.

The team at Waterloo adapted the radar system as a means to detect glucose levels in the blood non-invasively.

To do this, the system sends high-frequency radio waves into liquids that contain various levels of glucose and receive radio waves that are reflected back to it.

Then, the information on the reflected waves is converted into digital data for analysis by machine-learning AI algorithms.

“The data is processed on a server, and the results are near real-time, within a minute,” said Shaker.

The software can detect glucose changes on over 500 wave features or characteristics, including the length of time it takes for them to bounce back to the device.

The tests

So far, the researchers have tested the software with known controlled liquid samples and blood samples from volunteers at the Research Institute for Aging in Waterloo.

After comparing the data against standard glucometer devices, the researchers found that the results were 85 percent as accurate as traditional analysis.

“The correlation was actually amazing,” Shaker said in a statement. “We have shown it is possible to use radar to look into the blood to detect changes.”

The next step

Next, the researchers hope to refine the system to obtain results through the skin, which complicates the process due to conditions such as sweating or dryness, weight gain or loss, and the impact of weather.

They are also working with Infineon to shrink the radar device system to make it both low-cost and low-power.

Additionally, they are working to embed a wireless interface that can relay the data from a smartwatch to the cloud for data processing and analysis.

Currently, the data analyzed by AI algorithms is sent to computers, but the goal is to ultimately create a self-contained technology similar to smartwatches that monitor heart rate.

“We are working hard to speed up the development of a complete wearable prototype. We hope we can present something in the next five years,” said Shaker.