It’s every schoolboy’s dream: an easy way of looking through walls to spy on
neighbors, monitor siblings, and keep tabs on the sweet jar. And now a dream no
longer…
Researchers at the University of Utah say that the way radio signals vary in
a wireless network can reveal the movement of people behind closed doors. Joey
Wilson and Neal Patwari have developed a technique called variance-based radio
tomographic imaging that processes the signals to reveal signs of movement.
They’ve even tested the idea with a 34-node wireless network using the IEEE
802.15.4 wireless protocol, the protocol for personal area networks employed by
home automation services such as ZigBee.
Advertisement
The basic idea is straightforward. The signal strength at any point in a
network is the sum of all the paths the radio waves can take to get to the receiver.
Any change in the volume of space through which the signals pass, for example
caused by the movement of a person, makes the signal strength vary. So by
“interrogating” this volume of space with many signals, picked up by
multiple receivers, it is possible to build up a picture of the movement within
it.
This story is only available to subscribers.
Don’t settle for half the story.
Get paywall-free access to technology news for the here and now.
In tests with a 34-node network set up outside a standard living room,
Wilson and Patwari say they were able to locate moving objects in the room to
within a meter or so. That’s not bad, and the team says there is ample
potential for improvement by increasing accuracy while reducing the number of
nodes.
The advantage of this technique over others is, first, its cost. The nodes
in such a network are off-the-shelf and therefore cheap. Other through-wall
viewing systems cost in excess of $100,000. The second advantage is the ease
with which it can be set up. Wilson and Patwari say that adding a GPS receiver
to each node allows it to work out its own location, which should dramatically
speed up the imaging process. Other systems have to be “trained” to recognize
the environment.
Wilson and Patwari have even worked out how their system might be used:
“We envision a building imaging scenario similar to
the following. Emergency responders, military forces, or police
arrive at a scene where entry into a building is potentially
dangerous. They deploy radio sensors around (and potentially on top
of) the building area, either by throwing or launching them, or
dropping them while moving around the building. The nodes immediately
form a network and self-localize, perhaps using information about the
size and shape of the building from a database (eg Google maps) and
some known-location coordinates (eg using GPS). Then, nodes begin to
transmit, making signal strength measurements on links which cross
the building or area of interest. The received signal strength
measurements of each link are transmitted back to a base station and
used to estimate the positions of moving people and objects within
the building.”
That’s ambitious, but if they do get their system to the point where it can
be used like this, it raises another problem: privacy.
How might such cheap and easy-to-configure monitoring networks be used if
they become widely available? What’s to stop next door’s teenage brats from monitoring
your every move, or house thieves choosing their targets on the basis that
nobody is inside?
Of course, in the cat-and-mouse game of surveillance, it shouldn’t be too hard
to build a device that disables such a monitoring network. But only if you know
it’s there in the first place.
There are fun and games galore to be had with this idea.
Advertisement
Ref: arxiv.org/abs/0909.5417: Through-Wall Motion Tracking Using Variance-Based Radio Tomography Networks