Harnessing light for wireless communications | Hacker News Reader

Harnessing light for wireless communications

code.facebook.com

76 pointsdwaxe31 comments

Maybe with this technology, we can finally build the 80's neon cyberpunk future we've all been waiting for.

I wanted to upvote this twice.

One of the major challenges for FSO is that it requires direct line of sight (LoS). One reason that electromagnetic radiation in the RF range is more desirable for communications is because it can be non-Los (think about your cell phone, you do not have to see the tower for it to work, you can be indoors for example).

The second issue is the free space pathloss (see https://en.wikipedia.org/wiki/Free-space_path_loss ) you can see by the formula that the loss goes up exponentially by not only the distance but also the frequency, which is why the wireless carriers are so interested in the 600Mhz spectrum, you need to build fewer towers using 600Mhz than say 3.5Ghz.

The sudden interest in mm-wave (e.g. 28Ghz) for 5G is because the spectrum is much wider, that is the channels can be very wide allowing more data to be encoded, the trade off is that you start losing the NLoS capability, or it becomes much more difficult.

With light, all those problems are orders of magnitude more difficult to overcome. Also, there are regulations around the level of emissions of light for safety reasons, if a laser is turned up too high it can harm a persons eyes if they happen to look at it, so although the spectrum itself is not regulated, the amount of power you can emit is. That makes it very difficult to implement an FSO system that can safely transmit more than one or two kilometers, but if you get fog, rain or snow, that distance quickly becomes 100's of meters at best.

They are correct that the best way to overcome this issue is by developing sensors with much better sensitivity, by making them larger you can theoretically collect more photons, like a telescope. Since it is unlikely you can overcome the safety issue of increasing the transmission power level, it is smart to focus on the receive end and improve the sensitivity. But that does not help the LoS problem.

But a very interesting development nevertheless.

Loss goes up quadratically with distance, not exponentially.

The graphs showing the quadratic dependence of loss vs. frequency assume a half-wave antenna, meaning the antenna size decreases inversely to wavelength, so most of the effect is from having a smaller antenna.

In any case, path loss doesn't usually limit network capacity. Densely deployed systems are limited by interference, and the interference experiences the same path loss as the signal.

Yes, by the square, this is HN I cannot speak so loosely:) For Multiple access RF systems yes, but got point to point systems, that is line of sight, noise is not an issue, free space loss, also impairments from fog, trees and other things are dominant.

I think you are hugely overestimating the importance of non-line-of-sight communication.

Your example is cell phone communication. That is not the intended use case of this technology. Facebook's laser comms is part of their Aquila project, attempting to build a network of drones to supply internet.

This application works just fine with a line-of-sight constraint. At 50,000ft just about everything is line of sight.

Anyone building backhand RF links also hugely care about line-of-sight. Microwave towers carrying multi-gigabit speeds are carefully arranged for line of sight: this much improves reliability and throughput.

Yeah, was going to say something similar. UHF/VHF don't have "line of sight" but if I've got a hill between me and someone else I'm sure as hell going to have a hard time hearing him.

They put repeaters up on tall areas for that reason. Same with cell towers, yes you can get through non-dense material but there's still a signal loss associated with it.

The effect detailed in this comment and its parent is the Fresnel Zone [1], which isn't quite direct, human intuition 'line-of-sight', but that concept applied to RF waves.

Essentially, any object intruding into a blimp-shaped region between the receiver and the transmitter has the potential to mess things up, and is corrected by having a payload in the signal that's tolerant or self-correcting of errors.

[1] https://en.wikipedia.org/wiki/Fresnel_zone

Yes, if they are thinking about deploying vertically to a LEO satellite then perhaps LoS is not an issue unless you are indoors. But if you want to require the user install an antenna or receiver outdoors then it can work, but again, the atmospheric loss can be an issue, for example if it is cloudy outdoors then light will have difficulty getting through that. Some wavelengths may be better than others but anyone who has had satellite TV knows what happens when it rains. But I do think what they are doing is interesting and has a long term promise for a lot of use cases.

All of these challenges you've described are why Facebook's drone group also has a team working on 60/70/80 GHz PTP millimeter wave radio tech. Not very dissimilar from existing high-capacity 1GbE 60/70/80 GHz bridge radio products on the market now, from a radio perspective (example, one polarity in a 1500 MHz wide FDD channel in the 71-86 GHz bandplan, 256QAM modulation with ATPC+ACM), but self-aiming and motorized with stepper motors/servo mechanisms and service-channel data transfer for GPS-assisted aiming at a moving platform.

Yes, the higher you go in the frequency the more it is like light. But even at mm-wave such as say for example 28Ghz you still have penetration through walls, but the loss is very high, for 28Ghz it is about 19db for a according to IWPC for a typical house wood wall. For light, the loss is basically infinity. Building a user device, like a phone or tablet with stepper motors or servo mechanisms is probably not very realistic I don't think.

I'm not talking about anything other than going through free space air with clear Fresnel zone, this is for backhaul/trunk links... A drone might be a flying series of sector antennas/phased array in 5.x GHz serving CPE radios, but it also needs an uplink.

I don't really get why the range is this low. I can see flashes of light from a signal mirror from a place more than 10km away (and there were signal mirror links of ~100km in very good weather). In these systems, the noise should be way lower (there are many sources of relatively-low-frequency visual noise that I know of and I know of none that would generate at least hundreds of MHz frequency noise) and attenuation should be similar. Signal mirrors are eyesafe, so similar signal levels should be possible.

What am I wrong about or missing?

Good question, you can see a light flash on and off but to encode large amounts of information onto a channel is determined by the signal to noise ratio per Shannon's law https://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theore... the information rate is proportional to the SNR. So you may be able to see a light flash on and one once per second but to put 1Gbps on it you need a pretty high SINR. Things like dispersion come into play.

Do you mean dispersion as in "a pulse gets longer in time due to different length of paths taken by different photons due to scattering"? If not, what do you mean?

My point about seeing a flash was precisely that SNR is high, because I can distinguish the flash from noise.

EDIT: SNR depends on frequency only inasmuch noise level depends on frequency.

After reading this post, I was left with the question of why bother with the luminescence--why not just collect the transmitted light with the fibers and transmit it directly to the detector. The linked journal article makes it clear that the transmitted light is not collected by it entering the ends of the fibers. Instead, the transmitted light "scatters" off of the dye (changing wavelengths in the process), regardless of where and at what angle it hits the fiber, and some of it is scattered in the proper angular range to be transmitted along the fiber.

Wavelength shifting should make the detector more effective. Most photodiodes prefer longer-wavelength photos.

In addition, their goal with this project is to make it omni-directional, so capturing and coupling the incoming photons directly would be a challenge.

There may also be a little bit of wavelength-dependent trickery, where green photons are better at staying coupled, but I'd have to dig more to decide.

Efficiently directing light into a fiber optic cable is a difficult problem, in general. The luminescence is their proposed solution to this problem and one of the key results in their journal article.

You have to shine light end-on into a fiber--you can't couple light into the fiber by shining on it at the side. Even for large, visible-light scintillating fiber used in this paper, you only have a few degrees of possible incidence angles that direct the light into the fiber. See for example, (one of the first results from Google), the figure on page 4 of this datasheet [0].

[0] https://www.ethz.ch/content/dam/ethz/special-interest/phys/p...

Reminded me of the Ronja project: http://ronja.twibright.com/about.php

Isn't it amazing how we in the "developed" nations always presume that the "developing" nations need our technology? Colonial Europe assumed the savage nations needed railroads and hospitals, and now us enlightened 21st century liberals assume the developing world needs our internet, our network of connectivity.

But do they? Are they not already connected, and are we truly connected? Consider our societies. We are epidemically depressed, living alone, and becoming increasingly friendless. The internet has undoubtedly created business opportunity (I'm work remotely as a researcher); but has it "connected" us as we were not already connected? Instead, has it in fact disconnected us in its facade of relationships?

Do you mean places like, for example, Daru, Papua New Guinea? You can see this place from Australian territory.

I mean, we gave these people the OK Tedi mine disaster. There was a Cholera outbreak there around 2010-11. Electricity to run water pumps is available for about an hour a day.

I feel absolutely no qualms with thinking it'd be at least okay, possible even great, if everyone had something even near the opportunities we have here in Australia.

Also, your comment assumes a rose-tinted pasted that very well may not have existed. It wasn't that long ago (my grand parents era) where Polio was something we lived with. Where we really more connected 'back then'? Was domestic violence and alcoholism and less?

> (I'm work remotely as a researcher)

Sounds like you're projecting your own existential crisis onto others. Maybe you should try quitting your job and finding something local and more hands on.

No projection of my own. I have a significant person in my life and friends. I'm commenting on the well-documented reality of some western cultures. I few searches brought these stories up:

Depression: http://www.nimh.nih.gov/health/statistics/prevalence/major-d...

Living Alone: https://www.washingtonpost.com/politics/more-americans-livin...

Friends in UK: http://www.telegraph.co.uk/men/active/mens-health/11996473/2...

My immediate question was isn't their bundle of optical fibres performing essentially the same function as a lens? Why not use a lens?

Answering my own question, having just read the abstract (in case anyone else had the same reaction). The claim is that they get the light collecting powers, combined with a wide field of view. A lens collects the light, but has a narrow field of view, which requires steering/alignment.

Edit: Another possibility is a Luneburg lens, a type of lens that can focus plane waves from many directions to a small area? https://en.wikipedia.org/wiki/Luneburg_lens

> In addition, this portion of the spectrum is not regulated

Apart from not pointing a laser at a flying aircraft. I'm guessing this would be affected by weather.

Clouds and optical turbulence are both problems that have to be mitigated by FSO systems

I wonder what their detector's dark count rate and activation threshold is. If both are low enough, this is precisely the sort of detector people doing free space quantum crypto need.

Are they trying to re-invent Li-Fi ? https://en.wikipedia.org/wiki/Li-Fi

the article is easily digestible and you'll find that no, they aren't. not even close.

Hm, they seem similar because both Li-Fi and free-space optical communication (FSO) use visible or near-visible spectra to send and receive information. I guess they might be seen differently as Li-Fi is potentially designed for indoor use? I don't see why they are substantially different, however, or at least incomparable.

What the post discusses: https://en.wikipedia.org/wiki/Free-space_optical_communicati...

One significant difference is that FSO operates over a much longer distance and has to contend with atmospheric interference (clouds, turbulence, etc). Also, the transmitter/receiver hardware used is much different.

Reply on news.ycombinator.com