‘Invisibles’: The New AI-Enabled Wireless Monitoring Technology

In recent years ‘wearables’ have taken health care technology to new levels, allowing for real-time tracking of key physiological functions. Now, a research group at MIT has developed a new AI-enabled wireless monitoring technology known as ‘invisibles’. This smart technology offers all the benefits of ‘wearables’ without you having to wear anything on your body. Innovations such as these show the potential of artificial intelligence (AI) to transform health care.

Gain an introduction to wireless monitoring technology with Dina Katabi, Guest Expert in the Artificial Intelligence in Health Care online short course from the MIT Sloan School of Management.

Transcript

The role of machine learning in health care

Of course, there are many things that are changing in health care by introducing machine learning and AI, from understanding medical images to analyzing medical records of patients. But I am particularly interested in telling you about the advances that we have in digital health, and introducing digital health into our understanding of health care. And when I tell people about digital health, they typically think about wearables. They say, “Oh yeah, we are seeing these wearables such as Fitbits, for example.” That is used increasingly in understanding health and getting information about patients. I want to tell you about the next thing beyond the wearable, which is what we call ‘the invisibles’.

Discover new advances in digital health

We have invented a smart Wi-Fi-like box that we call the ’emerald box’. It’s very much similar to your Wi-Fi at home — it sits in the background of the home and it analyzes the electromagnetic waves in the environment. By analyzing these wireless signals in the environment, it can tell your breathing, your heartbeat, your movement, interaction with other people, without even asking you to wear any sensor on your body. Of course, when I tell people this, everyone is like, “Oh, like, how is that possible even? How can you get my breathing, my heartbeat without even touching me?” And it is really the power of these advances in machine learning, you should understand that, let’s step back: If you look around you, you are surrounded by a sea of Wi-Fi signal, of satellite signal, so there are plenty of wireless and electromagnetic waves around you. Now, every single move that you do, like you took a breath, the pulsing of your blood, it changes the electromagnetic shield around you. And what we are able to do is to design new machine learning algorithms that can analyze those changes in the electromagnetic waves to be able to tell.

I want to show you what we can achieve. Here, what you see on the slide, one of my students, this is our lab at MIT and one of the offices in the lab, and you see an arrow, that arrow is pointing to the wall because actually the device — our smart, wireless box — it’s not even in this office. It is in the adjacent office and it’s watching this person through the wall. Now, we’re going to make this person walk. And I want you to look at the red dot and how it’s able to track his motion. I want you to remember that it’s tracking him through the wall by just analyzing the wireless signal, and you can see that it tracks him pretty accurately.

Remember, like he has no sensors, no accelerometer, no cell phone on him, this is purely based on how wireless signal changes as it reflects off his body.