Hacker News new | past | comments | ask | show | jobs | submit login
5G is for sensing, not communication [pdf] (aalto.fi)
84 points by macawfish 6 days ago | hide | past | favorite | 35 comments





I submitted this article with the title:

"5G is for sensing, not communication"

It'll surely be changed at some point. I didn't think this would go to #1 within 20 minutes of my posting it. I chose the title out of frustration, to be frank, as a tongue in cheek protest of the stifled, polarized, repressed, sad state of public understanding of what RF technologies are capable of and our collective inability to have substantive conversations about these capabilities, let alone organize around issues of accountability and transparency in their deployment.

It's a reference to the article submitted a few days ago about a future Wi-Fi sensing standard [1]. I just saw that article, went to the comments and didn't see one single comment mentioning 5G.

Look at the last sentence of the abstract:

> Our results in realistic settings with five subjects suggest that accurate recognition of activities and environmental situations can be a reliable implicit service of future 5G installations.

My "academic speak" brain looks at this as a researcher essentially raising a huge flag, saying to us "hey, in case you weren't paying attention, we're making a collective decision to allow passive sensing technology everywhere".

I'd like to think we could have that kind of conversation before mm-wave beamforming antenna arrays were installed on every other street corner in the country, backed by proprietary hardware and software.

But at the rate things are going, we're on track to have the conversation after the fact. When all of our devices are hooked in, when we're dependent on the standards, when the consequences are immediately in front of us and our institutions are lined up to exploit the emerging applications.

1. https://news.ycombinator.com/item?id=29901587


Last few days on HN have been slow, bad luck!

As for 5G being used as a surveillance grid, I guess... But that only works in densely populated areas, where a few facial recognition cameras on every block would also do the trick (and that already exists, at least in the UK).


Think about any neighborhood dense enough for modern cell coverage. Then imagine a database that records all point to point EM anomalies. One could use such a database to determine when all persons leave their homes and where they went throughout the day, even without any person carrying a connected device.

We already have a system for that in the US. Just in the visible spectrum.

https://www.cato.org/blog/judge-allows-warrantless-aerial-su...


That should be "not just communication".

I agree, in hindsight

This article is only cited by 3 people according to ieee. Your frustration is probably just paranoia at this point.

Well then check out this whitepaper from the 5G 2021 Meeting:

"5G Localization and Context-Awareness"

https://www.5gitaly.eu/2018/wp-content/uploads/2019/01/5G-It...

Or how about this one, "5G Network-Based Passive Radar". I don't have access to the full text, but the abstract tells me enough:

> This paper presents the pioneering results of successful target detection using a fully operational and cooperative 5G cellular network as a source of illumination in a passive radar system.

https://ieeexplore.ieee.org/abstract/document/9661315

Or this one, a high level technical roadmap: "Enabling Joint Communication and Radar Sensing in Mobile Networks -- A Survey"

https://arxiv.org/abs/2006.07559

Another one: "High-Accuracy Localization for Assisted Living - 5G systems will turn multipath channels from foe to friend" (300+ citations)

https://wides.usc.edu/Updated_pdf/witrisal2016.pdf

There are at minimum dozens of peer reviewed articles about both active and passive sensing and localization applications of mm-wave antenna arrays of all sorts, including those involved in the 5G project. I can't list them all, but here are some keywords you can use in google scholar:

- device free localization

- passive localization

- perceptive mobile networks

- passive sensing

- passive identification

- joint communication and sensing

Just add "5G", "mobile networks" or "mm-wave antenna array" and you're good to go


5G is great for sensing but does a relatively poor job at what it's purported to do. The extremely high frequency means there are issues with range, obstacles like walls and even the weather. A cellular connection that get affected by rain isn't a smart idea. The advertised and theoretically high speeds aren't reached in the real world, unless...

>5G wavelengths have a range of about 1,000 feet, not even 2% of 4G's range. So to ensure a reliable 5G signal, there needs to be a lot of 5G cell towers and antennas everywhere. We're talking on every lamppost, traffic light, etc. because even trees can block 5G signals.

>Antonio Villas-Boas: 5G isn't gonna be cheap. You know, each node, or mini cell tower, needs some kind of connection to it, and that means laying down fiber optic cables, and, you know, it's still an undertaking, and it's definitely not in the millions. It's definitely in the billions, possibly hundreds of billions.

>Michelle: Not only will this cost billions of dollars, but there's also pushback from many local communities.

>Antonio: One of the biggest problems they face is actually local governments, local communities, who don't want these carriers to build towers or antennas all over the place. Or maybe they're afraid of the health risks, which is another big concern.

https://www.businessinsider.com/5g-high-speed-internet-cellu...


5G isn't just millimeter wave. Over the next few years, the majority of 5G will be on the 2496 to 2690 MHz (T-Mobile on recycled Sprint spectrum) and 3450 to 3980 MHz (AT&T, Dish and Verizon on 3.45 GHz, CBRS and C-band) bands.

>Over the next few years, the majority of 5G will be on the 2496 to 2690 MHz

Of course it will be because of the above mentioned problems. There is no way they'll be able to put up that many antennas any time soon, it will be a slow process. Those frequencies you cite are about the same as current 4G, with only the latter being slightly but not significantly higher. It's part of the reason people who try 5G end up disappointed it's not as fast as they expected.

The question is who is this actually a good deal for? Do consumers understand they're going to pay those hundreds of billions? Do they understand 5G means having antennas all over the place? And do they understand the privacy concerns? Most in my experience do not. I also find it interesting that right from the start there was a ton of noise in the news about "conspiracy theorists" fearing 5G for ridiculous nonsense reasons. Just the sort of thing you'd expect from a black propaganda campaign. But there was few talk about the actual tech, no talk about who decides these standards and why the entire world is supposed to sign up.

Do people know for example a Huawei guy is chairing the group that's bringing us wlan sensing with a noval 802.11bf wifi standard as well? So the milimeter scanners will be in our homes and outdoors as well. In my experience almost no one outside of niche tech bubbles knows anything about this. They just know faster internet good, they have no idea about the politics behind all this.

https://www.ieee802.org/11/Reports/tgbf_update.htm


It's worth pointing out that 5G is just a protocol which can, in principle, be operated on any of the frequency bands currently used for LTE. For instance, T-Mobile is currently providing 5G and LTE concurrently on their 600 MHz band.

> because 5G uses a totally different wavelength than the 4G standard your phone currently uses. The 5G standard uses millimeter waves, which are a lot shorter than the wavelengths 4G uses.

That article is just wrong there. Yes, 5G adds additional frequency bands, but no, those do not need to be used for 5G, and it also uses the same ones as LTE does. Which your usual 5G upgrade does.


I use 5G in lieu of fibre. I live about a half mile from the tower and I get 200Mbps consistently. Connections rarely drop, I am ssh’d into cloud machines for hours without issue. All this for £25pm and maybe £350 for the SIM unlocked 5G router. UK based.

Is this unmetered or you stay w/in the cap?

Unmetered

Not sure the 1000ft statement is accurate. I have 5G ISP and am ~4200 feet away and it works great: http://ewams.net/?date=2022/01/11&view=TMobile_5g_Internet_F...

That's because you're not on those frequencies, see the other comments. 4G(+) works great, sure. You're not downloading 20 Gbps, are you?

5G spectrum is not limited to high frequencies.

5G spectrum goes from high-bands 24-40 Ghz, mid-bands 1-6Ghz, and low bands < 1 Ghz


But aren't those hz bands limited in their throughput to about where 4g is?

increased synchronization between stations should allow for better beamforming to go along with the positional information, increasing throughput somewhat in the face of interference/attenuation

IIRC there is also a lot of improvement in channel multiplexing both at raw physical layer and link layers to allow e.g. numerous IoT type devices to have a low-rate cell connection without a plan.

This also allows packets to start at more times, which means that latency can decrease. (aka, lower ping times.)

Ah so a more robust and reliable, multi station system is formed? What's this beam forming and positional information stuff? Are they tailoring signals to individaul devices? That's so cool! This stuff is starting to feel like electro magnetic magic, haha.

yeah- they call it time-domain-multiplex (or something) when you have different phones subscribing to listen at different points in time. Traditionally only the closest station would broadcast to each phone, because multiple stations broadcasting the same frequency leads to a mushy signal (on average). BUT, if all the stations are synced, they can align the phases of their signals so that they 'constructively interfere' into a brighter signal where the target phone is. The appropriate phase difference will be different for each phone, so each time slice signal is indeed tailored to each device-- they also need to track device movement, to best target the signal while it drives/walks around.

'beamforming' is a kind of confusing term, based on how a bunch of antennas in a tight group can time their phases into a cone of signal-- functionally aiming the antenna, but without anything physically moving. When the involved antennas are not in a tight group, the 'beam' that is formed starts to look less like a cone, and more like a target 'bean'... beanforming ....

anyway this demo page can show some of how it works, ui is kind of weird, type a number and click to add/move an antenna, hit 'r' and click to place the receiving device, 'o' to align the phases for best signal.. https://apenwarr.ca/beamlab/


>"not communication"

Don't know about that part of the submitted title here.

There's plenty in the linked pdf about sensing, but I didn't read anything about it not being for communication. We've been using this rough part of the spectrum for both communications with microwave radios and for sensing with radar, for a long time.


This post has disappeared from the first 15 pages, in spite having 75 points and being only 4 hours old.

The title was not changed and it hasn't been flagged.

edit: I checked and it still appears under "new". It's weird that it's gone from the front page.

edit 2: I found it on page 27 :D


flagging isn't binary, and affects ranking before an explicit marker gets shown.

What used to seem so simple on the surface is inevitably more complex than I'd assumed. It makes sense though. I'm guessing this something that frequent HN posters come to understand sooner or later, and now I understand too.

I am sensing the title was not communicated well.

mmWave 5G is usable for sensing, but will not penetrate your home walls. I don't really see the issue.

TL;DR - 5G enables the kind of thing shown in the climax of The Dark Knight where every cell phone in the city is turned into an all seeing radar.

Title is misleading at best.

Lol, guy worries cell phone tech might be used for tracking…

That ship has sailed. They already know when you leave your house.


Perceptive mobile network applications don't require people to have phones on them to be tracked, identified and monitored. Many passive sensing applications are being researched. Some require additional hardware, others only data analysis. But all of them rely on the relatively high densities of mm-wave RF being rolled out with 5G deployments (not to mention the THz frequencies being proposed for "6G").

Here are some of the applications listed in the paper "Enabling Joint Communication and Radar Sensing in Mobile Networks - A Survey" (https://arxiv.org/pdf/2006.07559.pdf)

- (Through-the-wall) localization and tracking;

- Human behavior recognition and fall detection;

- Monitoring of biomedical signals such as respiration patterns;

- Human presence detection and radio fence

- Extensive on-street and open space surveillance for security and safety




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: