Hacker News Reader

Comment by ajsnigrutin | Hacker News Reader

People tend to forget, that if you can transmit 3km away, all the signals from everyone in that 3km circles are being received by your device too. With 2.4ghz wifi in busy apartment buildings, all the channels are overcrowded already, there is some space left on 5ghz, but not a lot, so that's why we've been going higher in frquencies for more bandwidth.

We also have other systems for unlicenced long-range communications (eg. LoRa), but once that becomes (more) popular, the spectrum will be full of that too.

View on HN

> everyone in that 3km circles are being received by your device too.

This could be pretty intense. Using the below tool I had a play and found that you could have about 1.1 million people within 3km of you when in Cairo.

https://www.tomforth.co.uk/circlepopulations/

I was thinking that number seemed crazy, as it is closer to ~10k at 3km for me(near Cambridge, UK). However, using that site and checking a few friends houses I see Cairo isn't that unusual; ~400k in Camden, 750k in Paris' 11th. And, making a couple of guesses at denser places led me to 1.7 million in Mumbai.

If anything it makes more amazed that regular wifi works as well it does between the house and garden, and I'll try to remember that next time I'm bemoaning its magic.

Assuming it’s mostly accurate (it has errors in my area), you can find bits of Cairo that exceed 2.1 million in a 3km circle, and Mumbai I can get over 1.8 million.

I’m in New Zealand and at that density you could fit our entire population in just over 2 of these 3Km circles.

I feel irritated when people stand too close to me, but at the Cairo density there must be tens of people within a stone throw.

The paper talks about how they can get it to work in a narrow 1MHz band (though at full 1MBbit/sec speeds I am skeptical of course), but even then I'm struggling to imagine fitting the number of people on that beach into channels that isn't a disaster. Strange to not address it...

Maybe if they formed a mesh network that prioritized signals of a certain power level and had dynamic power levels to avoid losing touch with the mesh while avoiding blocking close + strong signals.

Exactly, high absorption creates the localized cells each of which is typically limited in the number of (transmitting) devices. For something like this long range IoT cell to work you would want to create temporal cells or limit them to very low power. The problem is that those require agreement between devices (and noise sources) while simple absorption doesn't (treats everything equally).

Would dynamic directional RADAR be feasible? Or would overlapping RADAR signals going in different directions interfere with each other?

Given the passive radar code in KrakenSDR apparently fell afoul of ITAR, one would have to be careful around this.

Some routers do have 3/4/5 antenna directional antenna systems. I'm not sure if there's enough room or cost margin for an IoT device. If they were only directional then, I'm not sure they would help all that much, if there were other devices in the same direction (within 3km). However, there might be very short range differential (eg quadrapole) that could have faster than r^2 fall-off. The problem is that you're relying on all other devices in-band to do the same.

I think that's one of the reasons why unlicensed (ISM in the case of WiFi) spectrum is best used where the natural cell size (due to attenuation) is small.

It could make more sense if you have like a techie ranch in rural Wyoming and don’t have cell coverage I guess

Reply on news.ycombinator.com