Scientists Tweak AI System to See People through Solid Walls


A woman facing a wall. / Photo by: melpomen via 123RF


The ability to see through walls is a part of certain fictional powers like x-ray vision, but scientists at the Massachusetts Institute of Technology demonstrated that artificial intelligence can do it in reality.

The AI system they used in the study applied the innovation of Wi-Fi signals in recent years. Initially, the system can track movements through a solid surface like a wall using Wi-Fi, but the signals returning to the AI are inadequate. To work around it, the scientists created a machine learning algorithm to improve the returning signals and to enhance the accuracy of the AI in detecting subjects.

In a test involving the dubbed RF-Pose AI system, the new technology produced 3D stick-like figures that represented an individual detected through walls. The figures were shown with different colors that signified the heat signature of the person. Moreover, the AI was able to track down the pose of the subject.

“People have been trying to detect people through walls since the 70s. Around 2013, we showed that we can track people accurately. What’s new here is that, for the first time, we can create a dynamic skeleton of the person, their posture, and how they’re moving,” said Dina Katabi, the lead author of the study and a professor at the Computer Science and Artificial Intelligence Laboratory at MIT.

The system was able to perform all of these unimaginable tasks by transmitting signals through a Wi-Fi network and analyzing the patterns that bounce back to the network, quite similar to how echolocation works among bats and dolphins.

While the technology they developed is likely going to be used in surveillance, it also has other potential applications such as police safety, patient treatments with neurodegenerative disorders, and even video games.