3D Tracker Works Through Walls, Allows Gesture Control of Games

WiTrack tracks users as they move around a room (R) or lets them control computers by gesture (L).

Researchers at MIT have refined a system for tracking people on the other side of a wall, allowing that tracking to take place in real time and 3D.

The system, called WiTrack, locates people to within 10cm to 20cm using radio waves broadcast from a pair of WiTrack antennas, which reflect from human bodies and are picked up by a third antenna.

A previous version of the system, called WiVi, relied on ordinary WiFi signals, which are able to penetrate walls, but didn’t have the bandwidth to cover much distance or adjust when the person when upstairs or downstairs.

WiTrack uses radio frequencies 100 times lower than WiFi and structured in ways the receiver can recognize to accurately track the delay between when a signal is sent and it returns to the transmitter, according to Fadel Adib, a graduate student co-developing the system with MIT Computer Science and Artificial Intelligence Lab (CSAIL) researcher Dina Katabi.

WiVi improved on most radio-tracking systems, which require the tracked people to stand right in front of the transmitter most of the time, or carry a transponder with them. Instead, WiTrack builds a geometric model of the building, and the location of the person within it, based on the difference in lag time between the signals sent by each of two WiTrack antennas.

The system automatically screens out reflections from walls, furniture and other non-human objects “to allow us to focus on tracking human motion,” Adib said. (Video demonstration here.)

The system is accurate enough to locate a single body part and identify when it is moving, potentially allowing people to control computers by gesture, even when the person and computer are in different rooms.

WiTrack, which is able to identify when a person has fallen as well as when he or she has gone upstairs or down, could improve on tracking systems designed to safeguard the elderly by eliminating the need to carry transponders or call buttons, Adib added.

Camera-based tracking systems – such as Microsoft’s Kinect for Xbox One and Xbox 360 – use cameras to locate and identify individual humans. That is an effective way to locate a person in a room, but requires placing cameras in that room, and puts a lot of stress on the system analyzing the pictures to identify movement, according to Victor Bahl, director of mobility and networking research at Microsoft Research, who was quoted in the release.

The WiTrack system is able to track someone’s location on several floors, through walls, and without having any sensors or other objects in the room with the person being tracked, and does it accurately and without the computationally-demanding need to analyze video images to identify a location, Bahl said.

WiTrack can be used to locate the elderly, small children or others whose location needs to be monitored for safety reasons, but may have a richer future as part of a video-game system, in a manner similar to Kinect.

“Today, if you are playing a game with the Xbox Kinect or Nintendo Wii, you have to stand right in front of your gaming console, which limits the types of games you can play,” Katabi said. “Imagine playing an interactive video game that transforms your entire home into a virtual world. The game console tracks you as you run down real hallways away from video game enemies, or as you hide from other players behind couches and walls.”

Katabi and Adib introduced the WiFi-based version of the system, and an upgrade in October. The next step is to expand the system so it can track more than one person at a time, Adib said.

Adib and Katabi will demonstrate the current version at the Usenix Symposium on Networked Systems Design and Implementation in in Seattle, April 2-4, 2014.


Image: Katabi/Adib/MIT