MIT researchers use Wi-Fi to recognize people through walls | Hacker News Reader

MIT researchers use Wi-Fi to recognize people through walls

gizmodo.com

55 pointsthro123721 comments

All of a sudden, tin foil hats get a whole new meaning.

The problem with tinfoil hats is that it just makes your signature more unique. Tinfoil reflects radio waves better than skin and other clothing.

Unless you change up the shape/location of the tinfoil regularly, in which case the system will likely think you're a different person every time.

Is there something special about the WiFi spectrum that makes this work? Does it work with any ambient radio frequency noise?

My guess is that you need to have the kind of wavelength that meaningfully interacts with human features and body composition (i.e. something that scatters and reflects off of you, instead of being absorbed). It's not specific to the WiFi spectrum though, there are plenty of frequency ranges that meet these requirements.

The benefit of wifi spectrum is that the parts are cheap and you generally won't get in trouble for transmitting signals without a license, which is what their system does. (It's not passive.) Unfortunately the wifi spectrum isn't that great for through-wall imaging in general -- you're going to lose a lot of transmission power if the wall is composed of just about anything besides drywall, glass, or wood, on top of the massive loss of power you get by default since the received signal's power is inversely proportional to the distance raised to the fourth power.

Is there a range of frequencies that does better? Don't frequencies that go through walls easily also go through bodies easily, or are there frequencies where the body is particularly reflective w.r.t. concrete walls?

X-rays would do better, but you know, dangerous. On a less sarcastic note, yes, there are frequencies which would be better suited for a specific target material, but like Jach said, regulations and cost are what keeps this things in the WiFi spectrum. I think it also has a kind of "wow factor" when told to non-technical people.

Scanners designed for this purpose use X-rays and are in use by counter-terrorism and law enforcement units: http://www.pbs.org/wnet/need-to-know/opinion/google-street-v...

This is a different technology, those are not radars as far as I know, and I'm not sure it's possible to build x-ray radars due to low reflectance/scattering.

Is there a reason why very little information showed up from the legs? Are jeans just great at absorbing but not reflecting wifi signals?

The problem is that legs make poor backwards reflectors due to their shape, so the reflected signals aren't detected since they're sent elsewhere. Though the material can make a huge difference too. (e.g. wool and nylon can significantly reduce the reflectivity, as can the presence of wrinkles.) The best reflective part of the body is the torso (especially the back) since it's relatively flat.

So recognition doesn't work if it was calibrated while you were wearing different clothing? That kills some of the use cases.

The wall being made of reinforced concrete also kills a few use cases. But human detection isn't hopeless in general. You can do pretty well (even gender distinction) by having a physically based model of what you're expecting, e.g. humans, on the ground level instead of at some angle, within a certain range, potentially moving but you can detect heartbeats too, and there are models of how the human body behaves at rest and while moving. (Edit: the assumption of movement really helps a lot since you can diff samples and subtract out all the static bits. So even if clothing attenuates the signal by x%, x < 100, if the human's the only thing in the scene that's moving you're ok.)

(I'm actually impressed/jealous with what they accomplished since it was something I tried and failed to do alone for a school project based on http://ocw.mit.edu/resources/res-ll-003-build-a-small-radar-...)

We had a wifi radar at school. Was scary good. Only downside was that it jammed wifi for half the engineering building

Url changed from http://www.theverge.com/2015/10/28/9625636/rf-capture-mit-wi..., which points to this.

what happens when two person stand behind one another. keeping their bodies close to each other. Is there is any way to detect how many person are actually standing there?

I also would like to know the theoretical answer. The practical will be, I guess, "wait for them to start moving around".

Physics is crazy: http://web.media.mit.edu/~raskar/cornar/

> Physics is crazy

This is my new favourite response.

In theory I think it would be possible to detect this situation with SAR imaging techniques (and probably other techniques), you just need to have a large enough bandwidth in your transmission signal and be at a sufficient stand-off distance from the targets...

Reply on news.ycombinator.com