MIT Researchers Teach AI To Walk Through Walls

Jason Dorfman, MIT CSAIL

Scientists have already used AI to read minds and predict the future (sort of). Now, we can add X-ray vision to its growing list of superpowers. A team at MIT have built a tool that can see through solid objects and even identify people based on their gait alone.

The tool doesn't use X-rays, which would come with the slightly problematic side effect of showering nearby people with radiation. Instead, it relies on radio waves and utilizes the same physics as Wi-Fi. That is that wireless signals in Wi-Fi frequencies can pass through walls yet bounce off the human body. For this particular system, however, the team used radio waves thousands of times weaker than your typical Wi-Fi, reports Wired.

For MIT's Computer Science and Artificial Intelligence Laboratory's latest project, RF-Pose, Dina Katabi and her students collated thousands of clips of moving people recorded on a camera, reducing the people down to stick figures. These videos of stick figures were then fed into a neural network alongside corresponding radio signals to train the device how to recognize movement and gait from radio signals alone. After a while, the neural network was able to sense movement without the clips.

What surprised the researchers was how well the device was able to "see" movement from behind a solid object, like a wall. Essentially, it had X-ray vision – it was able to translate the mess of radio signals into movement even though it was never directly trained on information from the other side of the wall.

"If you think of the computer vision system as the teacher, this is a truly fascinating example of the student outperforming the teacher," MIT professor Antonio Torralba explained.

Not only was it able to detect movement, but it was also able to identify the person doing the moving 83 times out of 100. 

The team envisions a future where this type of technology can be used to improve the health of an aging society, helping to monitor the health and activity of elderly patients. 

"We've seen that monitoring patients' walking speed and ability to do basic activities on their own gives healthcare providers a window into their lives that they didn't have before, which could be meaningful for a whole range of diseases," Katabi said in a statement. "A key advantage of our approach is that patients do not have to wear sensors or remember to charge their devices."

Of course, there are plenty more sinister uses that come to mind, too.

"Just like how cellphones and Wi-Fi routers have become essential parts of today's households, I believe that wireless technologies like these will help power the homes of the future," Katabi added.

 

Comments

If you liked this story, you'll love these

This website uses cookies

This website uses cookies to improve user experience. By continuing to use our website you consent to all cookies in accordance with our cookie policy.