Revealed: Pentagon’s Plan to Defeat Russian and Chinese Radar With A.I.

[nadeemkhan110]

The Pentagon’s Defense Advanced Research Projects Agency (DARPA) is working on a new generation of electronic warfare systems that are based on artificial intelligence (A.I.). If the program were to prove a success, the new A.I.-driven systems would provide the United States military a way to counter evermore-capable Russian and Chinese radars.

“One of our programs at DARPA is taking a whole new approach to this problem, this is an effort we refer to as cognitive electronic warfare,” DARPA director, Dr. Arati Prabhakar, told the House Armed Services Committee’s Subcommittee on Emerging Threats and Capabilities on February 24. “We’re using artificial intelligence to learn in real-time what the adversaries’ radar is doing and then on-the-fly create a new jamming profile. That whole process of sensing, learning and adapting is going on continually.”

Current generation aircraft—including the stealthy Lockheed Martin F-22 and F-35—have a preprogrammed databank of enemy radar signals and jamming profiles stored in a threat library. But if those warplanes encounter a signal that has not previously been encountered, the system registers the threat as unknown—which means the aircraft is vulnerable to that threat.


“Today, when out aircraft go out on their missions, they’re loaded up with a set of jamming profiles—these are specific frequencies and waveforms that they can transmit in order to jam and disrupt an adversaries’ radar to protect themselves,” Prabhakar said. “Sometimes when they go out today, they encounter a new kind of frequency or different waveform—one that they’re not programmed for, that’s not in their library, and in a time of conflict, that would leave them exposed.”

During peacetime, the Pentagon usually deploys a signals intelligence aircraft like the RC-135V/W Rivet Joint to collect data on a new waveform. That data is then sent to a laboratory to be analyzed so that a new jamming profile can be created. Those new jamming profiles are then incorporated into a jet’s—F-22, F-35, F/A-18 or any other fighter—operational flight program updates. “Eventually, months—sometimes years—later our aircraft finally get the protection that they need against this new kind of radar signal,” Prabhakar said.


In the years prior to the digital revolution when radar waveforms were rarely altered, that slow process might have been adequate. In the current era where a new waveform can be created very quickly with minor software tweaks, the current process leaves American forces vulnerable. “That slow moving world is now gone,” Prabhakar said. “It’s not that hard to modify a radar system today. If you think about, the same technologies that have brought communications and the Internet to billions of people around the world, those are the same technologies that people are now using to modify radars.”

It’s a problem that has cropped up in many different regions around the world, Pabhakar said. Right now, the only U.S. combat aircraft that have some capacity to analyze enemy waveforms in real time are the Northrop Grumman EA-6B Prowler—which is still serving with the Marines—and the Navy’s Boeing EA-18G Growler. While both the Growler and Prowler have pre-programmed onboard threat libraries, both jets carry electronic warfare officers (EWO). Those EWOs can recognize and analyze the unknown enemy waveforms and—based on their experience—figure out a way to jam them in real time to an extent. However, it’s far from perfect because it relies purely on the skills of an individual EWO.

If DARPA’s new A.I.-based electronic warfare system works, it would save the Pentagon time, money and potentially even save the lives of aircrew if they encounter a new enemy surface-to-air missile system or fighter radar. “So what all of that means is that our aircraft in the future won’t have to wait weeks, months to years, but in real time, in the battlespace, they’ll be able to adapt and jam this new radar threat that they get.”
Source: http://nationalinterest.org/blog/th...ns-plan-defeat-russian-chinese-radar-ai-15357

 

[randomradio]

“We’re using artificial intelligence to learn in real-time what the adversaries’ radar is doing and then on-the-fly create a new jamming profile. That whole process of sensing, learning and adapting is going on continually.”

This is something the French are already doing, been doing it for over a decade.
@gambit @moon_light @That_engineer_guy

 

[moon_light]

This is something the French are already doing, been doing it for over a decade.
@gambit @moon_light @That_engineer_guy

What make you think the French have been doing that for decades? if you are referring to DRFM then many USA system have that too, and that is nothing new

 

[Olaf One-Brow]

What make you think the French have been doing that for decades? if you are referring to DRFM then many USA system have that too, and that is nothing new

The nearly 50 year old AN/ALQ-99 can do that.

It's not a new capability and hardly one the US doesn't already have, we're just refining it and looking for quicker, more effective means of doing so.

 

[randomradio]

What make you think the French have been doing that for decades?

Because that's how the Spectra operates. It creates jamming profiles on-the-fly even for unknown signals.

The nearly 50 year old AN/ALQ-99 can do that.

It's not a new capability and hardly one the US doesn't already have, we're just refining it and looking for quicker, more effective means of doing so.

The article does talk about it. But a lot of that is done manually, not by the computers.

What they are talking about is adding this capability on tactical fighters, not on specialized aircraft.

 

[shah1398]

Problem is that the more U rely on machines the more are the chances that they be disabled by an enemy even sitting miles away. In this digital age sometimes the analogue systems give damn good results (not that am I referring to revert back to such systems). U can not just leave everything to machines to decide and respond as in the end its human brain who is the most superior one.

 

[Parmanu]

Basically taking away atleast one human errror and a a human job as well.

 

[XenoEnsi-14]

U can not just leave everything to machines to decide and respond as in the end its human brain who is the most superior one.

Where did I put my keyboard?

 

[Oldman1]

Problem is that the more U rely on machines the more are the chances that they be disabled by an enemy even sitting miles away. In this digital age sometimes the analogue systems give damn good results (not that am I referring to revert back to such systems). U can not just leave everything to machines to decide and respond as in the end its human brain who is the most superior one.

Or turning against us.

 

[moon_light]

Because that's how the Spectra operates. It creates jamming profiles on-the-fly even for unknown signals.

How is that any difference from ALQ-187(v)2 , ALQ-131 EA PUP , Falcon edge , ALQ-211(V)9 , ALQ-214(V)3 , Spectra , ASQ-239

The article does talk about it. But a lot of that is done manually, not by the computers.

DRFM isnot operate manually

What they are talking about is adding this capability on tactical fighters, not on specialized aircraft.

DRFM is common even on tactical fighter jammer : ALQ-187(v)2 , ALQ-131 EA PUP , Falcon edge , ALQ-211(V)9 , ALQ-214(V)3 , Spectra , ASQ-239

 

[randomradio]

How is that any difference from ALQ-187(v)2 , ALQ-131 EA PUP , Falcon edge , ALQ-211(V)9 , ALQ-214(V)3 , Spectra , ASQ-239

DRFM isnot operate manually

DRFM is common even on tactical fighter jammer : ALQ-187(v)2 , ALQ-131 EA PUP , Falcon edge , ALQ-211(V)9 , ALQ-214(V)3 , Spectra , ASQ-239

I think you have issues with understanding what is being said.

As the article explains, if the signal is not in the databank, then the aircraft will be vulnerable to that threat. This is true for both the F-22 and F-35.

“Today, when out aircraft go out on their missions, they’re loaded up with a set of jamming profiles—these are specific frequencies and waveforms that they can transmit in order to jam and disrupt an adversaries’ radar to protect themselves,” Prabhakar said. “Sometimes when they go out today, they encounter a new kind of frequency or different waveform—one that they’re not programmed for, that’s not in their library, and in a time of conflict, that would leave them exposed.”

DRFM uses the built in library. If the signal is not present in the library, then it does nothing. The AI is meant to create jamming profiles for that unknown signal, which the Spectra does easily. That's the reason why the Rafale always gets the top marks in any evaluation.

Also, Spectra isn't a DRFM jammer, it can do that, but that's not its limit. It is an active cancellation system.

Basically what I'm saying is typical DRFM systems are obsolete.

 

[moon_light]

I think you have issues with understanding what is being said.

As the article explains, if the signal is not in the databank, then the aircraft will be vulnerable to that threat. This is true for both the F-22 and F-35.

“Today, when out aircraft go out on their missions, they’re loaded up with a set of jamming profiles—these are specific frequencies and waveforms that they can transmit in order to jam and disrupt an adversaries’ radar to protect themselves,” Prabhakar said. “Sometimes when they go out today, they encounter a new kind of frequency or different waveform—one that they’re not programmed for, that’s not in their library, and in a time of conflict, that would leave them exposed.”

DRFM uses the built in library. If the signal is not present in the library, then it does nothing. The AI is meant to create jamming profiles for that unknown signal, which the Spectra does easily. That's the reason why the Rafale always gets the top marks in any evaluation.

Also, Spectra isn't a DRFM jammer, it can do that, but that's not its limit. It is an active cancellation system.

Basically what I'm saying is typical DRFM systems are obsolete.

what you said is absolutely nonsense as usual. A DRFM system is designed to digitize an incoming RF input signal at a frequency andbandwidth necessary to adequately represent the signal, then reconstruct that RF signal when required. The most significant aspect of DRFM is that as a digital "duplicate" of the received signal, it is coherent with the source of the received signal.They can do that to a signal at a new frequency and waveform that isn't in their threat library. And a completely new signal that isn't in the library doesn't mean the jammer became useless either, even a simple repeater jammer can deal with new signal that isn't in its library. What the article talking about is actually the ability to detect what the enemy radar is doing and developing a jamming profile that adapting to that, for example jamming a CW radar will be different from jamming monopulse even if they work at the same frequency, so how to deal with a new form of tracking . And there is no evidence that Spectra can do any better than ASQ-239 or ALQ-214v3 or ALQ-131 in this aspect either since it has to rely on a threat library too. Active cancellation is simply too overrated, frequency hoping, PRF jitters and pulse compression can all be used to deal with it

 

[randomradio]

what you said is absolutely nonsense as usual. A DRFM system is designed to digitize an incoming RF input signal at a frequency andbandwidth necessary to adequately represent the signal, then reconstruct that RF signal when required. The most significant aspect of DRFM is that as a digital "duplicate" of the received signal, it is coherent with the source of the received signal.They can do that to a signal at a new frequency and waveform that isn't in their threat library. And a completely new signal that isn't in the library doesn't mean the jammer became useless either, even a simple repeater jammer can deal with new signal that isn't in its library. What the article talking about is actually the ability to detect what the enemy radar is doing and developing a jamming profile that adapting to that, for example jamming a CW radar will be different from jamming monopulse even if they work at the same frequency, so how to deal with a new form of tracking . And there is no evidence that Spectra can do any better than ASQ-239 or ALQ-214v3 or ALQ-131 in this aspect either since it has to rely on a threat library too. Active cancellation is simply too overrated, frequency hoping, PRF jitters and pulse compression can all be used to deal with it

You failed at reading and understanding the article and my posts. Read the article carefully again, it talks about waveforms and frequencies, not some new form of tracking. When you talk about pre-programmed libraries, you are talking about signals.

Read very, very carefully.
“Sometimes when they go out today, they encounter a new kind of frequency or different waveform—one that they’re not programmed for, that’s not in their library, and in a time of conflict, that would leave them exposed.”

 

[moon_light]

You failed at reading and understanding the article and my posts. Read the article carefully again, it talks about waveforms and frequencies, not some new form of tracking. When you talk about pre-programmed libraries, you are talking about signals.

Read very, very carefully.
“Sometimes when they go out today, they encounter a new kind of frequency or different waveform—one that they’re not programmed for, that’s not in their library, and in a time of conflict, that would leave them exposed.”

As long as the frequency is within the limit of Jammer frequency coverage ( normally the limit of it's TWT and receiver )then it can produce a fake replica signal even if doesnt recognize the signal as a threat radar within it's library ( this is true even for repeater jammer , and obviously the noise jammer wouldnot care whether the signal is in threat library or not , they only care about signal-noise ratio), if the signal is outside the limit of jammer frequency coverage then the jammer wont be able to detect , much less fooling the threat (this hold true for any kind of jammer ). If the waveform characteristics is so noise- clutter like that the jammer dont even know it is a radar then it wont start jamming automatically ( this hold true for any kind of deceptive jammers we have at the moment even Spectra and ASQ-239, or ALQ-214)
another problem is new waveform can also result in a new form of tracking that often mean the old jamming method is not effective anymore , this new USA project aim to give jammer ability to adapt to that.

 

[randomradio]

As long as the frequency is within the limit of Jammer frequency coverage ( normally the limit of it's TWT and receiver )then it can produce a fake replica signal even if doesnt recognize the signal as a threat radar within it's library ( this is true even for repeater jammer , and obviously the noise jammer wouldnot care whether the signal is in threat library or not , they only care about signal-noise ratio), if the signal is outside the limit of jammer frequency coverage then the jammer wont be able to detect , much less fooling the threat (this hold true for any kind of jammer ). If the waveform characteristics is so noise- clutter like that the jammer dont even know it is a radar then it wont start jamming automatically ( this hold true for any kind of deceptive jammers we have at the moment even Spectra and ASQ-239, or ALQ-214)
another problem is new waveform can also result in a new form of tracking that often mean the old jamming method is not effective anymore , this new USA project aim to give jammer ability to adapt to that.

Nobody is saying a jammer cannot do that. They are saying it has to be done manually now.