This Crazy MIT Device Knows When You’re Happy or Sad

Researchers Fadel Adib and Mingmin Zhao, and Professor Dina Katabi/Photo by Jason Dorman, MIT CSAIL

Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) mean it when they say they know how you feel. They’re behind a device called EQ-Radio, which can read a person’s emotions.

EQ-Radio doesn’t use on-body sensors or facial-recognition software to figure out if you’re happy or sad, but instead relies on a wireless signal that measures changes in breathing and heart rate. What’s more, it doesn’t even have to be in the room with you to do so.

Mingmin Zhao, a Ph.D. student at CSAIL, says EQ-Radio uses an approach developed last year to “zoom in on the signal reflected off the body of a particular person,” and ignore reflections from other people and objects.

After someone has spent 10 minutes with the EQ-Radio to establish baseline data about his or her neutral, happy, sad, angry, and excited states, EQ-Radio can study his or her waveforms, or heartbeat measurements, and determine emotion with 87 percent accuracy. Even on first exposure to someone, EQ-Radio can guess emotion with 70 percent accuracy.

For several years, Professor Dina Katabi, the project lead of EQ-Radio, has been working to use wireless devices to monitor health. She developed a company called Emerald, which is focused on making technology to detect and predict falls for the elderly.

“The next step,” Zhao says, “was to see if we could capture meaningful information about other human behavior that is invisible to the naked eye, like emotions.”

The researchers see EQ-Radio being used in entertainment, consumer behavior, and healthcare. One application, Zhao says, could be in smart homes.

“Smart homes could use information about your emotions to adjust the music or even make suggestions, like suggesting that you go outside for a walk if you’ve been sad for a few days,” Zhao says. “We believe a system like this could someday even be used to help monitor and diagnose conditions like depression and anxiety.”

But don’t get too excited, tech junkies. An emotion-reading Siri isn’t a reality just yet. The team will present its work at the Association of Computing Machinery’s International Conference on Mobile Computing and Networking next month, but there are no immediate plans for commercialization.

Learn more in this video from CSAIL: