DARPA Wants Wireless Devices That Can Blast Through the Noise

coondoggie writes "What if your wireless communications just absolutely, positively have to be heard above the din of other users or in the face of massive interference? That is the question at the heart of a new $150,000 challenge that will be thrown down in January by the scientists at DARPA as the agency detailed its Spectrum Challenge — a competition that aims to find developers who can create software-defined radio protocols that best use communication channels in the presence of other users and interfering signals."

Comments

Solution:

[Gaygirlie]

Mother-in-law.

Re:Solution:

[AchilleTalon]

This protocol doesn't support encryption.

Re:Solution:

[BronsCon]

You've never met my mother-in-law, then. She make the most cryptic remarks.

Re:Solution:

[Gaygirlie]

Can she punch a budget through the Senate?

Well, she can melt steel just by staring at it. And if she's in the mood for it she can have such a chilling effect on any gathering of people that your fart will freeze mid-air. I have no doubt she could sway the Senate if she felt like it.

Re:Solution:

[smittyoneeach]

Is she for hire? I'm envisioning a Kickstarter project, unleashing M-i-L on Harry Reid, who ends up looking like a Ramirez cartoon.

Re:Solution:

[slick7]

A set of Bose 501's or 1201's might do the trick.

It was developed...

[Anonymous Coward]

years ago. It's called Morse Code.
When all else fails, Morse Code gets through.

Re:It was developed...

[X0563511]

Not when the signal is below the noise floor.

Spread-spectrum should fix this. Hell, OLIVIA (slow as ass) can be used even when below the noise floor, provided you are listening for it.

Re:It was developed...

[Anonymous Coward]

DARPA really needs to go back and read Claude Shannon's work from the 1940's. Given a certain bandwidth, noise floor and power level, you can only stuff so many bits through the channel per unit time. You want more than that maximum? You need more power, or more bandwidth.

Next.

Re:It was developed...

[icebike]

Why should they read Claude, when what they are asking for is not anything to do with bandwidth or power, but simply competing transmissions, some perhaps hostile? The DARPA Spectrum Challenge places no explicit restrictions on bandwidth, and perhaps only portability restrictions on power. I rather suspect DARPA has already developed radio technologies that would make Mr Shannon's jaw drop.

Re:It was developed...

[SuricouRaven]

Shannon's mathematics are valid only so long as the assumptions behind them are true, and the main assumption is that there is *a* communications channel. They do not apply when you are using multiple independant or semi-independant channels, such as in a MIMO radio setup or using multipath communcation. The succesful candidate may well have so many antennas it resembles a sea urchin on a stick.

Re:It was developed...

[icebike]

Shannon's constraints on the carrying capacity of an information channel are a fundamental part of information theory, not related to any given technology.

And also unrelated to the challenge under discussion, as I pointed out.

The suggestion by the GP that DARPA is unaware of Shannon, or that they are asking for a solution that somehow violates his constraints is simple wrong, and ill informed pedantry.

Don't worry, I have the solution here.

[Anonymous Coward]

You gotta use multi-phasic shifting!
Make it so.

Re:Don't worry, I have the solution here.

[AK Marc]

PSK has been around longer than you. Most of the tech words used in Star Trek were stolen from real technical terms that most ACs were too dumb to know. Though a winning solution would likely include some phase component (and other things that sound like sci-fi terms).

Re:Don't worry, I have the solution here.

[SuricouRaven]

My favorite was 'gravimetric sheer.'

It's a deliberatly technical name for what the common people call 'tides.' There was a strange tendency for starships to get nearly pulled apart by it, which suggests either those ships are a lot bigger than they look or someone (Q?) has been going around stuffing planets with small black holes as a joke.

Re:Don't worry, I have the solution here.

[AK Marc]

Since they had artificial gravity and could control gravitons, one could guess there were non-mass generated gravitational fields causing issues. How well would a ship hold together if every other deck reversed ship gravity?

Re:Don't worry, I have the solution here.

[Anomalyst]

Make it so.

cross-stitch or buttonhole?

Re:Don't worry, I have the solution here.

[SuricouRaven]

The show carefully avoids going into detail about how this artificial gravity works. Some episodes reveal that there is a point in every ship where the gravity is effectively zero, and as the ship isn't being constantly accelerated it could be concluded that the field is a zero-sum process: Half the ship does have gravity reversed compared to the other half. Presumably the turbolifts flip over when you cross the transition.

Or it could be that the artificial gravity exists purely because filming apparent weightlessness costs a fortune, and the writers don't really care about the technology involved. Star trek has never been hard science fiction - the fantastic settings, alien cultures and futuristic technology are used at their best to explore current issues, taken to a place where they can be examined from the perspective of an outsider.

Re:Don't worry, I have the solution here.

[AK Marc]

Most descriptions of artificial gravity make it sound like the deck itself send out gravity waves "up" to pull things towards them, like a plane of lighting approximates a point source infinitely far away (like the sun). And there's no reason that would accelerate the ship. It's being accelerated "up" like the earth is being accelerated into my feet. Sure, that's the theory, but it doesn't feel like that in reality. The deck is attracting itself to everything above it as if it had the mass of a planet 3000 miles away (you don't want the point source approximated in the deck itself, or your feet would weigh pounds while your head was weightless). So everything is attracted to it. But you are making the classic sci-fi mistake that if the effect isn't explained, it's because it's not practical. If you can create gravitons for something so trivial as ship gravity, and you have inertial dampeners and such, why couldn't you emit a countering field a the edge of the ship? Wouldn't that be standard practice? So people could never track you by your mass, you appear "weightless" to everyone. Otherwise, if someone stops in interstellar space, you could drop a bomb or leave mines and they'd make their way to the ship eventually. Seems pretty self-explanatory to me. They generate local gravity, and cancel it at the hull. That's consistent with the fiction (technology and every story I can think of).

direct link

[Trepidity]

The actual DARPA page, with rules/etc., is here.

Ideas?

[hhawk]

I'm not an engineer.. but some random thoughts?

1) Can you use out of band communications to trigger a special mode of the router (assuming you control it)? (e.g., A special listening mode that gives certain devices priority)

2) Rather than use typical QAM type of modulation, can you use a more limited constellation but BOOST the power so you can punch it through the noise? (which would also allow you to make very good use of forward error correction (FEC); again assumes that you are able to program both the sender and the router.

Assuming you can't alter the Network Stack on either device than you have to look at the communications itself

3) Other than that, it would seem you would need to use a programmable antenna/software antenna, Etc. turning your own Antenna into a high directional Antenna with as much gain as possible. Basically find the WIFI router you want (geo locate it, perhaps triangulation with the help of friendly nearby devices) and the push all your signal towards (a dynamic Yagi antenna that auto-magically maintains it's "aim" at the router even as the sender and/or the receiver move about.

4) If you did have some control over the router and you were sure that you could depreciate other traffic then the router would also get the same auto-magical antenna system and the two devices would "focus" on each other.

Re:Ideas?

[Gaygirlie]

1) Can you use out of band communications to trigger a special mode of the router (assuming you control it)? (e.g., A special listening mode that gives certain devices priority)

This isn't about routing the data, this is about getting the data to reach the other end at all. Besides, routing priority has already been handled by QoS for years now.

2) Rather than use typical QAM type of modulation, can you use a more limited constellation but BOOST the power so you can punch it through the noise? (which would also allow you to make very good use of forward error correction (FEC); again assumes that you are able to program both the sender and the router.

Assuming you can't alter the Network Stack on either device than you have to look at the communications itself

3) Other than that, it would seem you would need to use a programmable antenna/software antenna, Etc. turning your own Antenna into a high directional Antenna with as much gain as possible. Basically find the WIFI router you want (geo locate it, perhaps triangulation with the help of friendly nearby devices) and the push all your signal towards (a dynamic Yagi antenna that auto-magically maintains it's "aim" at the router even as the sender and/or the receiver move about.

Just increasing the power until the signal is heard is one way of doing it, sure, but there is always a limit to how high you can go. Plus then you're contributing yourself to the noise and broadcasting your location on the battlefield to everyone around you.

Re:Ideas?

[jovius]

I've been thinking that could it be possible to use wifi signals to have a radar. The packets carry a unique identifier, and if those packets come from multitude of sources, different directions and in different phases it should be possible?

Re:I think that it depends...

[TheLink]

If they don't care if it interferes with such devices, then isn't it just a matter of increasing the power output on their transmitter?

Yep, nuke it from orbit. That definitely sends a signal that'll blast through the noise...

Smoke signals :)

[youn]

Enough said :)

RTFM

[interiot]

The goal is to "engineer software-based radios that transmit data faster than a competitor using identical hardware".

The goal isn't to develop fancy new hardware, or to use an overwhelming amount of power. The goal is to develop fancy new software.

With frequency-hopping and time-hopping techniques, if you can intelligently adapt to the local interference, and transmit in the time and frequency gaps where the interference doesn't occur, then you can transmit more data for the same amount of power. That's the goal.

Re:RTFM

[Pinky's Brain]

The problem of course is that you don't necessarily know for certain the space of the gaps ...and by the time you detect a collision you already fucked up part of the aggregate bandwidth (or in the case of UWB raised the noise floor). There is no protocol agnostic solution to that problem.

I imagine the final implementation will be massively ad-hoc with a huge number of specialized algorithms for each part of the spectrum specifically adapted to the common protocols for that spectrum to make sure what you think are gaps are actually gaps.

They don't want just ANY solution

[davidwr]

They want the BEST solution.

Seriously, as others have pointed out, there are obvious ways to solve the problem.

DARPA is challenging the community to come up with optimal or at least more-optimal-than-everyone-else ways.

The only "downside" I can see is that the competition is in a controlled environment. I say let's do the tests in an "uncontrolled" environment, much like an outdoor sporting event where neither the competitors nor the officials get to control the weather. Play the "game" several times in several "representative" environments such as large cities, small towns, rural areas, and during several different times of day or night, etc. Then use a pre-determined, pre-published scoring rule to determine not only the "overall winner" but the "winners" in various categories, e.g. "best in big cities," "best at night," etc.

If achieved, the end of spectrum licensing?

[Neil_Brown]

If spectrum licensing is predicated on the basis of a need to prevent / minimise interference, if such a technology is developed, the requirement to license spectrum (and for governments to print money carrying out such licensing) would seem to fall away.

Yochai Benkler has already made a persuasive case (I don't know if this was officially published) around this and, if it was possible to deploy widely technology that worked irrespective of interference, we'd seem to be one step closer.

The cynic in me thinks it might fail as a result, since I doubt many governments would want to lose the money, or incumbent operators a means of excluding others from the market.

Re:If achieved, the end of spectrum licensing?

[AK Marc]

No. Because the point of this is the ability to get a signal through in an RF-unfriendly environment. The solution will have down-sides (higher power than optimal, lower bandwidth than optimal). So id everyone did this in shared spectrum, then we'd end up with lower bandwidth and higher power usage. We can do that now in 2.4. The point of licensed spectrum is that it's optimal. The noise floor should not change, and the bandwidth will be higher. The shared combative solution is a tragedy of the commons situation. The more who use it, the less there is to share.

Finding a place to put data

[Animats]

The rules aren't available on the site yet, but I assume they're interested in resistance to jamming. From a theoretical perspective, as long as the receiver isn't saturated, there should be some data rate at which transmission is possible. This follows from Shannon. Noise can be overcome with redundancy, at the cost of data rate.

You can usually do better than that by moving around the spectrum to quieter areas. That's what frequency-hopping systems do. Jammers can be agile too, but unless the jammer is in a direct line between sender and receiver, the jammer is always at a time disadvantage due to speed of light lag. Very fast frequency hopping can overcome agile jammers.

What DARPA wants, I suspect, are systems that package up all this into a system that takes care of any noise problems automatically and will get a message through if it is physically possible. DoD has had systems for that for decades, but the technology tended to assume that the opposition didn't know the details of how it worked. It may be possible to have jam-resistant systems that work even if the opposition knows the technology.

Re:Finding a place to put data

[canadiannomad]

DoD has had systems for that for decades, but the technology tended to assume that the opposition didn't know the details of how it worked. It may be possible to have jam-resistant systems that work even if the opposition knows the technology.

Now that is an interesting idea, maybe link the jumping to cryptographic keys, you have to know the private key in order to be listening to the right spot on the band at the right time. That way the jamming would have to know the keys in order to predict where it has to jam...

Yeah, talking out of my ass, but sounds like an interesting idea.

Re:Finding a place to put data

[AK Marc]

It would seem to me that frequency-hopping spread spectrum with predictive noise calculations would work. Hop *before* the noise moved. Use wide-band so that a single-frequency jam won't work.

(not the solution for the question asked here, but:I would look at sending a masking signal. One powerful transmission, with spurious emissions that look to be standard "shoulders" associated with over-driving an amplifier. However, the powerful emission is a one-time-pad key, followed by a random number identified by that key. The "shoulders" are separate transmissions designed to look like part of the main transmission, and are, in fact the actual signal. The receiver can't separate out the shoulders because they are overlapped by the larger signal, but the intended receiver has a matching OTP for the random number, and subtracts the expected decoy from the transmission, revealing the intended transmissions in decodable clarity. Sort of RF stenography. But that's a great solution that doesn't solve the problem asked here, nor does it hide the second signals if the enemy knows what to look for.

What they are asking for is "How do we guarantee drones are unjamable, even if the enemy has found intact crashed drones (all hardware is known) and the enemy jammer is 10m from the drone and our transmit station is 1000km away - making the ground jam (effectively) infinitely more powerful?"

For that, I'd say "use lasers". But then, that's the same as "use a different frequency" And the enemy would come up with a way to blast the drone with light.

Re:Finding a place to put data

[SuricouRaven]

The jammer could just jam every frequency at once. The use of software-defined radio would complicate the task though, as high-power wideband transmitters are tricky things to build compared to your basic narrow transmitter. The attacker would probably be best of finding an old-school radar engineer who knows how to build a few magnatons. They are simple enough to build in a garage (Getting the air out is the trickiest part), but can transmit a ridiculous amount of power and over a very wide spectrum. That's how microwave ovens are so good at jamming wireless LANs.

Re:Finding a place to put data

[Animats]

maybe link the jumping to cryptographic keys, you have to know the private key in order to be listening to the right spot on the band at the right time. That way the jamming would have to know the keys in order to predict where it has to jam...

That's how classical frequency-hopping systems work.

One of the interesting possibilities of doing this with a software-defined radio is to have the receiver listen to the whole band and recognize the signal of interest without knowing in advance what the transmitter will do. This avoids the shared key problem and the cryptosync problem. The transmitter's hopping pattern can be random, instead of psuedorandom, so the jammer can't predict it even in theory.

Re:Finding a place to put data

[Animats]

The jammer could just jam every frequency at once.

That's called a "denial jammer". Historically, denial jammers aren't very effective other than at very short ranges. It takes huge amounts of power to jam a whole band so solidly that no narrow-band redundant signal can get through. Denial jammers are very easy to find in peacetime and targets for homing anti-radiation missiles in wartime.

Re:Finding a place to put data

[SuricouRaven]

Short range can be plenty, if your only aim is to carry out an ambush and buy a few minutes of extra time. As for missiles... oh, that's just asking for some fun. Simply ask a group of local children to play somewhere near the jamming device. If the US forces do attack it, they'll only end up killing children, which is sure to rally some more support for your cause.

I'm thinking of Iraq-like insurgency scenarios: High-tech US forces with far superior equipment, but against an enemy much more familiar with the local culture and in urban areas where a high civilian population rules out the 'bomb first, ask later' approach. I could throw together a short-range denial jammer for the 2.4GHz band with ease, but jamming anywhere else would be harder due to the difficulty of obtaining an appropriate magnetron.

Re:Finding a place to put data

[Blaskowicz]

Although, the wars we're currently fighting are against an enemy whose most advanced communication system is a HAM radio. They wouldn't stand much to lose from denial jamming.

Really? The enemy uses satellite phones, maybe wifi 5.5 GHz links, or even a frequency hopping software defined radio - the design is here for all to see with the help of GNU Radio. Strong encryption is a possiblity (I don't think the NSA can even break your ssh or the https for your webmail, and those come for free).

So, the enemy may get some quite advanced technologies by ordering them on ebay or tinkering with things. Even deploy some Wimax or LTE-Advanced but granted, many of these things are probably very easy to jam - saturate their well-known spectrums. But, is it so easy to do for wide bands of multi-GHz frequency, and at a long range, I don't know. We also have wars that are remotely fought - Pakistan, Lybia, Syria, so no big trucks on the ground to use the jamming toys. I guess the "enemy" is often free to use satellite phone and even GPS. But all that stuff would have to be properly studied, these thoughts come from my ass.

Re:I think that it depends...

[icebike]

The only probable solution would be to dedicate specific channels for their use and have rigidly enforced laws in place which forbid usage by consumer devices.

I'm not sure how you rigidly enforce laws in battle field situations. If you could, why not just make a law against the enemy carrying weapons?

There are two key requirements in the Darpa Challenge:

1) High priority radios in the military and civilian sectors must be able to operate regardless of the ambient electromagnetic environment, to avoid disruption of communications and potential loss of life.
2) Response operations, such as disaster relief, further motivate the desire for multiple radio networks to effectively and efficiently share the spectrum without requiring direct coordination or spectrum preplanning.

In the end I suspect that the winning entry will pay little heed to the regulatory frequency allocations, and fall back on the FCC standard of non-interference by instantaneously finding unused frequencies over a wide spectrum and pushing messages through those spaces in small encrypted bursts so short that licensed users of that bandwidth would not even notice it. Alternatively you might be able to embed your transmission within already widely used frequencies for digital television, AM /FM radio, by using the gaps between allocated frequencies.

Re:I think that it depends...

[AK Marc]

If he's a non sequitur, then you are an off-topic.

Re:I think that it depends...

[davester666]

Yes, and it conveys 1 bit on information:

I don't want to talk to you anymore.

Throat Mics

[Jane Q. Public]

I don't know about the radio spectrum interference, but if they want relative freedom from sound interference, all they have to do is use something like the old WWII-style throat microphones that pilots used. They work. And of course I assume that the technology has gotten better since... they'd probably work even better now.

Re:Not reallly needed

[AK Marc]

The bandwidth isn't "licensed" so there isn't play nice going on. It's reserved. If you had an emergency with 5 states responding, they'll be saturated now. 4.9 isn't "unlicensed" like 2.4 is, but it isn't licensed (in the sense of one user "owns" it), either.

That and I read the question as "how do you guarantee drones will be unjammable, even if the drone hardware has been captured and the jammers are 5 orders of magnitude closer to the drone than the intended transmitter?" In which case, your answer is unrelated to the problem they want solved.

PPM

[thygate]

PPM (pulse position modulation) seems to be pretty resilient. Picking up mode-S transponder signals from planes as far away as 200km with a stock DVB-T antenna and a really cheap SDR (rtlsdr), which has very poor dynamic range. Even when the signal barely gets above the noise threshold.

150.000 is not a lot

[houghi]

If I can place my order just a few milliseconds faster, I will be able to make a LOT of money. If I make it possible,. how much do you think banks will be willing to pay me? 150.000USD will be pocket change.

I am sure I can at some zeroes when I offer it to the banks.

Re:I think that it depends...

[SuricouRaven]

The 'Turn up to eleven' approach could work for stationary equipment, but it isn't practical for things like field radios - which are exactly the sort of thing that a hostile force may want to jam prior to and during an ambush, to keep the targets from calling for backup.

Re:Not reallly needed

[SuricouRaven]

You work on making the drones work without communications. How hard can it be to make a program for 'Fly here. Shoot missile here. Return here.'?

An application from FM

[Chewbacon]

Frequency Modulation has a phenomenon known as capture. Whichever is the loudest is what you pick up in your car. So if you broadcast it with enough energy, it will break through all sorts of interfering signals on the same frequency.

Re:An application from FM

[jovius]

That's the reason why radio stations want to be as loud as possible, and use overwhelming dynamic compression on anything. The result mimics noise where continuous power is practically resting at 0db. That combined with already dynamically limited source material means that the output is total garbage of sound waves.

But it works, so...

Re:An application from FM

[MiG82au]

Loudness is unrelated to the amplitude of the carrier signal that's being frequency shifted. Hence the FM rather than AM. Besides, since when do radio stations have competing signals on the same frequency (do you know what capture effect is)?

Direct link is not really populated yet...

[girlinatrainingbra]

Re: The actual DARPA page, with rules/etc., is here.
.
Except for the fact that right now (~1pm PST 2012-12-22) that page has grayed out tabs for
1 -- Rules
2 -- Register
3 -- Q&A
;>)
The only tab that is live currently is the "Home" tab at that direct link http://www.darpa.mil/spectrumchallenge/. I don't think the other tabs go live until January 2013, so the rules and such are not available yet.

Re:Direct link is not really populated yet...

[icebike]

Send them an email. They will either laugh you into the delete bin, or send you a link to those pages.

Re:Direct link is not really populated yet...

[girlinatrainingbra]

There is a link at the bottom: a "mailto" link that you can send the correct subject line to, and they'll email you when the contest starts. They can't send a link to pages that aren't live or up yet. If you want to send an email, go ahead and do it. :>)

Re:Not reallly needed

[AK Marc]

The problem with that is they like to use the drones in changing situations. If you tell it to shoot something, that something may change in the time it takes the drone to get there. With remotely-controlled drones, you hover in the area, then shoot on demand, and remain to observe or return for re-arm as the situation dictates.

null steering

[sonoronos]

Like most DARPA competitions, this is not for novel research into techniques. The solution is usually selected from available technologies. I'm assuming that the most likely candidate is MIMO and null steering.

Re:I think that it depends...

[Anomalyst]

this was proposed a number of years ago. watchdng the noise floor over a wide swath of frequencies and using an algorythm to select the one used

Awesome

[jvillain]

I was wondering how the terrorists were going to trigger their IEDs when the growlers were flying over head. DARPA to the rescue.

Potential (ab)use of this technology

[fufufang]

I have a feeling that people might use this research to blast their data packet on the existing crowded frequencies.

This is a bit misleading...

[bwalzer]

The stuff about "disaster relief" is entirely bogus. They are asking people to work on an entirely military application for very little gain...

Re:This is a bit misleading...

[beinsvein]

The stuff about "disaster relief" is entirely bogus. They are asking people to work on an entirely military application for very little gain...

Granted, the pay is measly but the result could be used anywhere, not just in military applications. DARPA challenges generally let the participants keep their ipr, so if someone comes up with something good, expect civilian benefits. Noisy transmission channels is not a uniquely military problem.

Re:I think that it depends...

[mark-t]

How is it off topic? It was a suggestion that *IF* they didn't want their priority communications to interfere with normal consumer electronics, then there would probably have to be rigidly enforced regulations in place that would specifically prohibit those frequencies for use in consumers electronics.

The responder commented that laws prohibiting the use of those frequencies wouldn't do any good in battle field situations, which is entirely irrelevant to the point I was actually making.

So how am I off topic, exactly?