<i>> This proof-of-concept would be a breakthrough for healthcare, security, gaming (VR), and a host of other industries.</i><p>Similar capability is scheduled for new consumer routers in 2024 via Wi-Fi 7 Sensing / IEEE 802.11bf. Hundreds of previous papers include terms like these:<p><pre><code> human-to-human interaction recognition
device-free human activity recognition
occupant activity recognition in smart offices
emotion sensing via wireless channel data
CSI learning for gait biometric sensing
sleep monitoring from afar
human breath status via commodity wifi
device-free crowd sensing
</code></pre>
Earlier discussion: <a href="https://news.ycombinator.com/item?id=34423395" rel="nofollow">https://news.ycombinator.com/item?id=34423395</a><p>Sample code exists for ESP32 WROOM, <a href="https://wrlab.github.io/Wi-ESP/" rel="nofollow">https://wrlab.github.io/Wi-ESP/</a> and Intel 5300, <a href="https://dhalperi.github.io/linux-80211n-csitool/" rel="nofollow">https://dhalperi.github.io/linux-80211n-csitool/</a>
A variation of this was used in "The Dark Knight", and you'll recall that Morgan Freeman quit (ok or maybe threatened to quit per the comment below, we know he was back in The Dark Knight Rises) over the ethical implications.
The paper trains a model that a. requires training data in a specific environment. Meaning, to deploy this tech, you need access to the space first to train this model. b. once trained, will not transfer its knowledge if routers are moved, or setup is deployed in a new area. c. training was done with just a single person moving. multiple people moving was not evaluated by this tech/probably harder to achieve. With that in mind, I think privacy-invasion capability of this technology is exaggerated by some comments here.
Pretty cool. Semi-related but reminds me of this research from MIT on seeing around corners:<p><a href="https://youtu.be/JWDocXPy-iQ" rel="nofollow">https://youtu.be/JWDocXPy-iQ</a>
> In addition, they protect individuals’ privacy and the required equipment can be bought at a reasonable price.<p>To argue that this protects people's privacy (versus cameras in public spaces) is certainly a very odd take.<p>I'd be more curious to know what are legitimate use cases of this and who funded the research.
We have done the same and published back in early 2020. THis paper does reference our work as [11]. Blog post with our human detection video here (eerie !):<p><a href="https://futurebnd.com/human-detection-classification-and-visualization-using-wi-fi/" rel="nofollow">https://futurebnd.com/human-detection-classification-and-vis...</a><p>and the paper here:<p><a href="https://arxiv.org/pdf/2001.05842.pdf" rel="nofollow">https://arxiv.org/pdf/2001.05842.pdf</a>
I would really like to see a practical open source implementation of this so people can start looking for ways to defeat it. Otherwise you know the police are going to abuse this.
"WiFi Routers Estimates 3D Pose of Humans in Modelled Reconstruction" as an actual non bullshit title. The paper being a more concise "DensePose from WiFi".
The photo in article is some random stock image and has nothing to do with the research paper[1].<p>[1] <a href="https://vpnoverview.com/wp-content/uploads/2301.00250.pdf" rel="nofollow">https://vpnoverview.com/wp-content/uploads/2301.00250.pdf</a>
Slightly related,a little dated, but relevant reminder:<p>Rebehold, the Range-R:
<a href="https://en.m.wikipedia.org/wiki/Range-R" rel="nofollow">https://en.m.wikipedia.org/wiki/Range-R</a><p><a href="https://www.usatoday.com/story/news/2015/01/19/police-radar-see-through-walls/22007615/" rel="nofollow">https://www.usatoday.com/story/news/2015/01/19/police-radar-...</a>
Pretty similar to an older Nature paper, using smart speakers to monitor heart rate.<p>Wang, A., Nguyen, D., Sridhar, A.R. et al. Using smart speakers to contactlessly monitor heart rhythms. Commun Biol 4, 319 (2021). <a href="https://doi.org/10.1038/s42003-021-01824-9" rel="nofollow">https://doi.org/10.1038/s42003-021-01824-9</a>
Is there something low-cost and purpose-built to do this at short range, for when you just want to do some body tracking for VR/metaverse stuff? The trouble with doing this with a webcam is that occlusion results in errors and the avatar's limbs jump around. This sort of RF thing has more field of view, too.
I used to get excited about the benefical uses of new tech. Now, I'm expecting to be charged 19.95 subscription (for data that should never and doesn't really need to leave the LAN) for the privliege of amazon tracking my movements so they know when to send TP JIT. Thank you guys
It's not new as you could think after reading the article. I remember reading about this technology - seeing humans through walls basing on internal wifi signal for some US government agencies - at least 7y ago. Also OpenWRT software for routers has plugin for gestures for MANY years already. Possibilities of reading gestures by your home router implies all mentioned use cases.<p>It's past, but written here as a future. It's truism but written as discovery.<p>Another truism would be saying that it scales for any radio antennas - like GSM. Especially if you have dense grid.
I really wish we would make more of WiFi routers, AVM with their Fritzboxes seem to be the only ones doing something interesting, with features like SIP Server, SAT>IP Server, Internet Radio etc.<p>Routers can have attached storage, why can't we backup our phones to them instead of iCloud (instead of workarounds like Nextcloud)? Why can't they directly provide location information over a standard protocol, thus potentially avoiding a request to a cloud service?
I remember seeing a similar study years back… the claim was tracking of positions of bodies. The mechanism reminded me of Project Soli, using signal processing to make 2D/3D inferences. This sounds like the next iteration on that finding. Wild.
Link directly to the research paper:
<a href="https://vpnoverview.com/wp-content/uploads/2301.00250.pdf" rel="nofollow">https://vpnoverview.com/wp-content/uploads/2301.00250.pdf</a>
> This proof-of-concept would be a breakthrough for healthcare, security, gaming (VR), and a host of other industries.<p>They forgot the panopticon and Big Brother. Big breakthrough here as well!<p>Enjoy the week people, funky times ahead :)
Can't wait for the day when those scientists start coming up with systems that intelligently assess whether your wifi, phone, or even just electricity wires are being hijacked by malicious actors.
MIT CSAIL had an extremely similar project way back in 2015:<p><a href="http://rfcapture.csail.mit.edu/" rel="nofollow">http://rfcapture.csail.mit.edu/</a>
The more unexpected discoveries we find, the more I think how many more capabilities of everyday devices are "hiding" in plainsight that would surprise us.
You may be working with people who, at the beginning of their careers, worked on exciting and challenging projects as junior engineers for U.S. defense contractors to either detect the precise location of specific Wi-Fi clients.<p>Ask them when they realized that their work was extensible to any radio frequency client (cell, Bluetooth) and used for targetting missile strikes. I can guarantee you know at least a few people in the industry who did.<p>Just because we can doesn't mean we should. This story reeks of DoD funded research which somehow gets whitewashed as "cool new tech thing!" on tech blogs when it should really be sending chills down your collective spines.<p>This capability may be fringe and nation-state controlled for a few years, then it will inevitably fall into the hands of large and well-funded criminal organizations, abusive spouses, and of course overfunded trigger happy SWAT teams — who will still manage to get their court order addresses wrong and kill innocent people and pets over a no-knock warrant.<p>All this triggers in me is the irrespressible urge to get technologists to finally get it through their thick skulls that what we do <i>does</i> kill people exactly like doctors. We've just refuse to take responsibility for it when any other industry would have seriously discussed ethics board and licensure at this point. No matter how complicated such an effort would be.
<i>This proof-of-concept would be a breakthrough for healthcare, security, gaming (VR), and a host of other industries.</i><p>/facepalm<p>They do acknowledge the privacy concerns but go to make (imho) pie-in-the-sky arguments like 'this will enhance privacy because security cameras won't be as necessary in public spaces. Journalism doesn't pay much, so maybe this is some naively idealistic person's first writing job. I once believed that adding public comments on news websites would elevate the standard of public discourse and I mentally kick myself on the regular for the time I spent promoting this idea back in the 1990s.<p>The researchers offering the same ideas in the paper don't have such an excuse; they're creating an entire new class of surveillance technology and pretending that this will somehow enhance privacy, which flies in the face of all experience and research on the topic. The technicals result are outstanding and I'm very impressed by them, as well as the exposition and direction of research. The potential applications are numerous and exciting to my inner geek.<p>But I'm also worried. The existing limitations will fall sooner than expected, and it will be productized while the ethicists are still drafting their arguments (at which point they'll shift to asking for donations to counter the latest threat). Semi-seriously considering repainting the inside of my house to make a faraday cage by mixing copper paint in the underlayer.