Communications of the ACM

traverse the wall, reflect off objects and humans, and come back imprinted with a signature of what is inside a closed room. To convince yourself that Wi-Fi signals traverse walls, just recall how you can receive Wi-Fi from another room.

The main challenge of using Wi-Fi signals to see through a wall is that the wall’s reflection is very powerful. In fact, the wall’s reflection is 10,000– 100,000 times stronger than any reflection coming from behind the wall. As a result, the wall’s reflection will overwhelm the Wi-Fi device and prevent it from detecting any minute reflection coming from behind it. This behavior is analogous to how someone looking at the sun cannot see an airplane in the sky at the same time. The sun’s light would overwhelm the person’s eyes and prevent them from seeing the airplane, just as the wall’s reflection would overwhelm the Wi-Fi receiver and prevent it from detecting reflections from behind it.

To overcome this problem, the authors of this paper leverage recent advances in MIMO (multiple-input, multiple-output) communications. In MIMO, multiple antenna systems can encode their transmissions so that the signal is nulled (that is, sums up to zero) at a particular receive antenna. MIMO systems use this capability to eliminate the interference of unwanted receivers. In contrast, this paper proposes the use of nulling to eliminate reflections from static objects, including the wall. By eliminating the wall’s reflection, the proposed system can start registering the minute reflections from behind it. It analyzes these reflections to coarsely track the motion of a person behind a wall and count the number of people in a closed room.

Gesture Recognition with Wi-Fi

Pu, Q., Gupta, S., Gollakota, S., Patel, S.

Whole-home gesture recognition using wireless signals. ACM MobiCom, 2013; https://homes. cs.washington.edu/~gshyam/Papers/wisee.pdf.

This paper takes Wi-Fi-based motion
tracking to another level: it shows how
to use Wi-Fi reflections to recognize
human gestures. Specifically, over the
past few years there has been a grow-
ing interest in gesture-based user
interfaces. Past gesture-based inter-
faces, however, required the person
either to be directly in front of a sen-
sor (like the Xbox Kinect) or to wear or
carry a device (such as Nintendo Wii).
In contrast, this paper shows how to
perform gesture recognition through-
out an entire home without requiring
the user to hold or wear any sensor. It
does so by relying on Wi-Fi signals.

To capture information about gestures using wireless signals, this research relies on the Doppler effect. The canonical example of Doppler is the pitch of an ambulance siren that increases as it gets closer and decreases as it moves farther away. The authors leverage this concept using Wi-Fi signals.

In particular, Wi-Fi signals are transmitted at a carrier frequency (around 2.4GHz). A forward movement causes a small increase in this frequency (by a few hertz) and a backward movement causes a small decrease in this frequency. The authors observe that human gestures are typically composed of forward-backward movements. By zooming in on the frequency changes in the reflected signal and decomposing them into small movements, they show how to recognize human gestures. They use this capability to enable users to control appliances throughout their homes by performing in-air gestures.

Monitoring Breathing and Heart Rate Using Wireless Signals

Adib, F., Mao, H., Kabelac, Z., Katabi, D., Miller, R.C.

Smart homes that monitor breathing and heart rate. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, 2015, 837-846; http://witrack.csail. mit.edu/vitalradio/content/vitalradio-paper.pdf.

The final paper in this selection shows that we can capture and monitor human breathing and heart rates by relying on wireless reflections off the human body. To do so, the authors exploit the fact that wireless signals are affected by any movement in the environment, including chest movements caused by breathing and bodily movements caused by heartbeats.

The main challenge in extracting
these minute movements is that they
are easily overwhelmed by any other
sources of motion in the environment.
To overcome this challenge, the paper
first localizes each user in the envi-
ronment, then zooms in on the signal
reflected from each user and analyzes
variations in the user’s reflection to
extract breathing and heart rate. By
isolating a user’s reflection, it effec-
tively eliminates other sources of in-
terference, including noise or extrane-
ous motion in the environment, which
may otherwise mask the minute varia-
tions caused by the user’s vital signs.
This allows multiple users’ breathing
and heart rates to be monitored using
wireless signals, even if the users are
behind a wall.

Where Do We Go from Here?

These papers offer a few instances of a broader set of functionalities that future wireless networks will provide. These networks will likely expand beyond communications and deliver services such as indoor localization, sensing, and control. The papers presented here demonstrate advanced forms of wireless-based sensing to track humans, capture their gestures, and monitor their vital signs even when they do not carry a wireless device. This area of research is still nascent, and only time will tell how much further these techniques can go.

Fadel Adib is an assistant professor at the MIT Media Lab. He works on wireless networks and sensing systems. His research has been identified as one of the 50 ways MIT has transformed computer science over the past 50 years. The BBC, NBC, CBS, the Washington Post, the Boston Globe, and The Guardian have covered his work.