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

[gambit]

DRFM uses the built in library. If the signal is not present in the library, then it does nothing.

I hope you are not saying that the 'memory' in DRFM means a library.

Memory here means to keep a sample of the signal in memory to work on it. A DRFM system can use a library and nothing else. In that case, if a signal memory is not recognized via the library, then the system will do nothing, or at best tell the operator that it does not recognize the signal.

An adaptive DRFM is what SPECTRA claimed to be and for all purposes, it is what it claimed to be.

But the problem for any DRFM system is when the signal's pulse characteristics are variable at a rate that is beyond the DRFM system's sampling rate.

For the interested laymen...

The above is a simple illustration of the foundation of most radar signal: the pulse.

Any of the pulse's elements can be variable by itself or in combination with the others. The more variable, the more difficult to analyze.

The above is a sample of the common variations on a series of pulses. Next is the creation of a 'pulse train'. Just like the locomotive, where there is a start and an end, a radar pulse train have a start and an end.

Assume a DRFM system that has a library of known signatures but does not rely solely upon that library, and that this system is also adaptive.

Given the fact that this DRFM system cannot know the exact physical length of any incoming radar signal, it must have a finite sampling rate time. Simply put, it cannot ingest the incoming signal until pulse train termination. What if the pulse train is a few millions pulses ? Or tens of millions ? The more complex the pulse train's characteristics, the greater the memory burden. There has to be a limit on how long can this adaptive DRFM system remains in 'listening' mode and start doing 'something'.

If the hostile radar signal came from a sophisticated system, read 'First world' category, pulse train characteristics will vary from train to train. The idea is not new as some weather radars already have limited capability of creating variable pulse trains to distinguish different types of hydrometeors ( rain, snow, sleet ) in a complex weather phenomena. Different polarization is one method. So in facing a hostile radar that can vary its signals from pulse train to pulse train, an adaptive DRFM system can be overwhelmed if those variations are outside of the DRFM system's sampling capability.

 

[randomradio]

There has to be a limit on how long can this adaptive DRFM system remains in 'listening' mode and start doing 'something'.

As you already know, radar signals travel much farther than the radar's ability to detect a small target. The Rafale uses that time before it is detected to process the signals and localize them.

If the signal appears to be from the enemy, say the pulse train is jittered or staggered or is using dwell/switch PRIs, that's the best case for the Rafale because Spectra knows it is a dangerous signal. In this case, Spectra stays in listening mode only for 1/2 wavelength of the first pulse. It doesn't wait to study the waveform, it reacts instantly. But for this to be effective, it must know for sure that the signal is hostile which is not particularly difficult to figure out in most cases.

Of course, nothing bad will happen if Spectra attacks false alarms, but that's a waste of precious transmitter resources. Currently, Spectra is equipped only with a few dozen T/R modules for this purpose. But by 2021, Rafale will be upgraded with over 2000 or 3000 additional T/R modules, which will help Spectra deal with far more threats at once.

As for the processing of the signal itself, whether it is 100 pulses or 1 billion pulses, Spectra can theoretically be in listening mode forever. The signal can be processed using multiple sampling frequencies if real time data is required, but that's independent from the jamming. Memory isn't a roadblock anymore, but it's good to have as much processing as possible if you want your jammer to react on the first pulse.

 

[moon_light]

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

So how is Spectra is any difference ? if it isnot recognized the signal as radar signal then there will be no jamming automatically, same as any other jammers

If the signal appears to be from the enemy, say the pulse train is jittered or staggered or is using dwell/switch PRIs, that's the best case for the Rafale because Spectra knows it is a dangerous signal

it the signal is jittered and change frequency randomly then how can you know it is a threat? it would be pretty much like back ground noise ( since they also change randomly )

. In this case, Spectra stays in listening mode only for 1/2 wavelength of the first pulse. It doesn't wait to study the waveform, it reacts instantly. But for this to be effective, it must know for sure that the signal is hostile which is not particularly difficult to figure out in most cases.

how can it be easy if the signal isn't even in the threat library ?

. But by 2021, Rafale will be upgraded with over 2000 or 3000 additional T/R modules, which will help Spectra deal with far more threats at once.

the whole nose of Rafale have less than 1000 T/R modules and ESA cant scan pass 60 degrees boardside, so where do we put the 3000 T/R modules ?

 

[randomradio]

So how is Spectra is any difference ? if it isnot recognized the signal as radar signal then there will be no jamming automatically, same as any other jammers

Spectra uses algorithms that allow it to generate a real time database of any unknown signal and even use it immediately. It was built that way since the late 90s. Other systems are nowhere as versatile, at least the DARPA director confirms that when it comes to both the F-22 and the F-35.

I do not have any information beyond that.

In fact, stuff that you talk about and is being used on jets today has been tested in France back in the 90s.

it the signal is jittered and change frequency randomly then how can you know it is a threat? it would be pretty much like back ground noise ( since they also change randomly )

This signal is localized very accurately. Plus this data is combined with other data that comes from other sensors, including radar. So it won't behave like a simple RWR with very inaccurate direction finding ability.

how can it be easy if the signal isn't even in the threat library ?

That's the thing about Spectra. It attacks pulses, it doesn't need to know the full pulse train.

But there is advantage to know the pulse train beforehand because it helps make far more accurate predictions.

A DRFM jammer attacks the radar, Spectra attacks the signal itself.

the whole nose of Rafale have less than 1000 T/R modules and ESA cant scan pass 60 degrees boardside, so where do we put the 3000 T/R modules ?

You know something? In 2021, the Rafale's nose won't carry the radar. The radar itself will be moved to everywhere else on the fuselage. And this system won't be restricted to the X band.

 

[moon_light]

Spectra uses algorithms that allow it to generate a real time database of any unknown signal and even use it immediately.

That is literally the definition of DRFM, it records unknown signal to use it immediately. But that is not the same as ability to recognize signal that isn't in threat library. Nevertheless that is completely different from what described by DARPA
“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.”

Source: https://defence.pk/threads/revealed-pentagon’s-plan-to-defeat-russian-and-chinese-radar-with-a-i.446237/#ixzz4IpCALrxL
Ability to learn what enemy radar is doing and develop a jamming profile that suitable to that is an enormous jump in capability

Other systems are nowhere as versatile, at least the DARPA director confirms that when it comes to both the F-22 and the F-35.

Neither F-22, F-35 or Rafale have the capabilities to recognize signal that isn't in its library as threat and jamming automatically at the moment

I do not have any information beyond that.

In fact, stuff that you talk about and is being used on jets today has been tested in France back in the 90s.

Even the old ALQ-165 have DRFM, so it really isn't something new

This signal is localized very accurately. Plus this data is combined with other data that comes from other sensors, including radar. So it won't behave like a simple RWR with very inaccurate direction finding ability.

ASQ-239, ALR-94, ASQ-213, ALQ-218 all have very good direction finding - geolocating capabilities, and sensor fusion isn't something that only Rafale have, it common among US fighters too

That's the thing about Spectra. It attacks pulses, it doesn't need to know the full pulse train.

Technically even old repeater jammer can generate jamming signal halfway of the pulse, but the problem is if you don't have the whole pulse train, your jamming signal will be rather easy for radar to recognize from real reflection

But there is advantage to know the pulse train beforehand because it helps make far more accurate predictions.

Actually, unless you know the pulse train before hand, active cancellation would be quite ineffective

A DRFM jammer attacks the radar, Spectra attacks the signal itself.

That is nonsense

You know something? In 2021, the Rafale's nose won't carry the radar. The radar itself will be moved to everywhere else on the fuselage. And this system won't be restricted to the X band.

ESA array cant scan pass 60 degrees board side, so even if you put the arrays on the fuselage there is only quite specific place you can put them (and rafale will also gain a significant amount of weight if you decide to do so, not to mention heating and power problem
Btw where is the source for your statement that Rafale will have 3000 T/r modules by 2021?

 

[randomradio]

That is literally the definition of DRFM, it records unknown signal to use it immediately. But that is not the same as ability to recognize signal that isn't in threat library. Nevertheless that is completely different from what described by DARPA
“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.”

Source: https://defence.pk/threads/revealed-pentagon’s-plan-to-defeat-russian-and-chinese-radar-with-a-i.446237/#ixzz4IpCALrxL
Ability to learn what enemy radar is doing and develop a jamming profile that suitable to that is an enormous jump in capability

I never said it cannot have additional capabilities.

Neither F-22, F-35 or Rafale have the capabilities to recognize signal that isn't in its library as threat and jamming automatically at the moment

Sorry, but you can remove Rafale from that list. It does.

ASQ-239, ALR-94, ASQ-213, ALQ-218 all have very good direction finding - geolocating capabilities, and sensor fusion isn't something that only Rafale have, it common among US fighters too

The F-22 has sensor fusion, but only in the RF domain. The others don't. F-35's capabilities are yet to be fully developed.

Actually, unless you know the pulse train before hand, active cancellation would be quite ineffective

No.

That is nonsense

Haha. But it is true. Spectra starts transmission of the exact pulse it receives, but out of phase by 1/2 wavelength. Waiting for the pulse train to terminate or even finish 30% is irrelevant to Spectra for active cancellation.

The benefit of processing the waveform is only in order to quickly identify the threat.

ESA array cant scan pass 60 degrees board side, so even if you put the arrays on the fuselage there is only quite specific place you can put them (and rafale will also gain a significant amount of weight if you decide to do so, not to mention heating and power problem

There is no such problem. It will use GaN which is 10 times smaller than GaAs. And with half the weight. Cooling requirement is many times lower.

Btw where is the source for your statement that Rafale will have 3000 T/r modules by 2021?

This is how the theory works, but this design won't be used because it is not as effective. This presentation slide was made in 2007, it's not relevant now, but helps explain how it works.

In this design, the aircraft still has a radar in the nose. But in 2021, the nose will not carry a radar. Anyway, this design alone has over 3000 modules. And this is designed primarily for GaAs systems. GaAs has far too much size and weight constraints.

GaN can be built around the fuselage. Like this-

So it can be curved around the airframe as necessary. Which means there is no need to use the radar in the nose anymore. And the surface area of the fuselage is very high, significantly higher compared to the nose.

 

[moon_light]

I never said it cannot have additional capabilities.

you still dont understand as usual, the capability to record signal to build a library isnot something new, ISR assets do that all the time, it is completely different from what the new adaptive jamming here can do, which is developing a jamming profile that is most suitable to the current threat even if such threat isn't in the recorded library yet

Sorry, but you can remove Rafale from that list. It does.

No it doesn't, sorry but your wishful thinking doesn't make it true, there is no source or evidence even suggest such thing

The F-22 has sensor fusion, but only in the RF domain

because it doesn't have IR sensors

. The others don't.

F-18E/F and F-16E/F does

F-35's capabilities are yet to be fully developed.

Already reached IOC, pilot already praise it's sensor fusion

No.

Yes it is whether you want to admit it or not, Active cancellation rely on destructive interference so without knowing exactly what they transmit before they do, your jammer will always leave the begin of the pulse untouched ( because it not gonna process faster than speed of light), An if you leave the begin of the pulse untouched the jamming can be filter out easy by leading edge tracking ( similar way to filter out RGPO)

Haha. But it is true. Spectra starts transmission of the exact pulse it receives, but out of phase by 1/2 wavelength. Waiting for the pulse train to terminate or even finish 30% is irrelevant to Spectra for active cancellation.

I understand active cancellation, you dont have to re explain it. The problem with yoir theory is if adversary use pulse compression then your cancelation signal will end up amplified the reflection. Secondly you will end up trying to produce and transmit "cancellation signal" to everything from background clutter to datalink and IFF, and side lobes, even the ground reflection of enemy's radar

The benefit of processing the waveform is only in order to quickly identify the threat.

NO, the benefit of processing waveform is so that you dont transmit randomly and everywhere, not only ineffective but also make you vulnerable to HoJ missiles

There is no such problem. It will use GaN which is 10 times smaller than GaAs. And with half the weight. Cooling requirement is many times lower.

No, GaN is not 10 times smaller than GaAs ( and you cant just shrink down the antenna to what ever size you like unless you want your gain become ridiculously low) , GaN has higher out put and power density, but if you want to transmit higher power, you will need more cooling, energy cant be destroyed.
And yes ESA cant scan pass 60 degrees boardside because they need wave interference to work

This is how the theory works, but this design won't be used because it is not as effective. This presentation slide was made in 2007, it's not relevant now, but helps explain how it works.

In this design, the aircraft still has a radar in the nose

That is no where even close to 3000 T/R modules even for X band let alone a wide range of frequency, F-22 and F-35 both have a bunch of receivers at wing edge, you can get away with 10-20 T/R at most, the wing isn't that thick

. But in 2021, the nose will not carry a radar.

So where is your sourceto support statement that Rafale will have 3000 T/r modules by 2021?

Anyway, this design alone has over 3000 modules. And this is designed primarily for GaAs systems. GaAs has far too much size and weight constraints

That is not only horrendously ridiculous but also stupid, physics problem doesn't just disappeared because you have a new material. More transmit power will require more cooling there is no way around it.
And you will also need a lot more generator to supply your 3000 T/R modules. And no matter how light your modules is, it will be much heavier than a simple composite or metal airframe

GaN can be built around the fuselage. Like this-

So it can be curved around the airframe as necessary. Which means there is no need to use the radar in the nose anymore. And the surface area of the fuselage is very high, significantly higher compared to the nose.

GaAs can be built like that too, it called a smart skin, has nothing to do with the material. And you will still have weight, cooling and power problem

 

[randomradio]

you still dont understand as usual, the capability to record signal to build a library isnot something new, ISR assets do that all the time, it is completely different from what the new adaptive jamming here can do, which is developing a jamming profile that is most suitable to the current threat even if such threat isn't in the recorded library yet

Sorry, but you haven't understood the OP.

No it doesn't, sorry but your wishful thinking doesn't make it true, there is no source or evidence even suggest such thing

Even the manufacturer website says it.

http://www.dassault-aviation.com/en/defense/rafale/the-sheer-power-of-multisensor-data-fusion/

Overcoming individual sensor limitations related to wavelength / frequency, field of regard, angular and distance resolution, etc, by sharing track information received from all the sensors,

No aircraft does this.

F-18E/F and F-16E/F does

They don't. There are only four aircraft known to have sensor fusion, Mirage-2000, F-35, Rafale and F-22.

What the other aircraft do is they simply use sensor data which is most reliable to declutter the display. That's not sensor fusion. This psuedo-fusion happens post processing.

On the Rafale, all the sensors send raw data to a single computer. And all that data is crunched to create a single air picture without dismissing data from other sensors. Rafale has pre-processing sensor fusion.

Already reached IOC, pilot already praise it's sensor fusion

The most important capability is not even close to ready. The top tester himself says it.

http://www.bloomberg.com/news/artic...till-falls-short-pentagon-s-chief-tester-says

The most complex software capabilities “are just being added” and new problems requiring fixes and verification testing “continue to be discovered at a substantial rate,” Gilmore wrote to Air Force Secretary Deborah Lee James; General David Goldfein, the service’s chief of staff; and Frank Kendall, the Pentagon’s acquisitions chief.

Check the article.

I understand active cancellation, you dont have to re explain it. The problem with yoir theory is if adversary use pulse compression then your cancelation signal will end up amplified the reflection. Secondly you will end up trying to produce and transmit "cancellation signal" to everything from background clutter to datalink and IFF, and side lobes, even the ground reflection of enemy's radar

The Rafale carries plenty of sensors to avoid that situation. Pulse compression doesn't affect active cancellation. Localization of the signals is 0.1deg.

No, GaN is not 10 times smaller than GaAs ( and you cant just shrink down the antenna to what ever size you like unless you want your gain become ridiculously low) , GaN has higher out put and power density, but if you want to transmit higher power, you will need more cooling, energy cant be destroyed.
And yes ESA cant scan pass 60 degrees boardside because they need wave interference to work

Nobody is shrinking the antennas.

GaN needs lesser cooling than GaAs. Safe operating temperatures are 2.5 times higher. GaN burns at 400degC, GaAs maxes out at less than 200degC.

That is no where even close to 3000 T/R modules even for X band let alone a wide range of frequency, F-22 and F-35 both have a bunch of receivers at wing edge, you can get away with 10-20 T/R at most, the wing isn't that thick

There are 10-20 on the canard leading edge alone, for Spectra. The ones on the wings are much larger. But we need to wait for the layout. The design in the image is generic.

So where is your sourceto support statement that Rafale will have 3000 T/r modules by 2021?

A Mirage-2000 pilot told me the date. There are various news articles also explaining about conformal arrays.

And you need 3000 for the system to be effective. Take the image as an example, the main array has 1000, the two side arrays will have 500 each, that's 2000 right there. The remaining 4 arrays could have 200-300 each, that's 800-1200. Total = 2800-3200.

That is not only horrendously ridiculous but also stupid, physics problem doesn't just disappeared because you have a new material. More transmit power will require more cooling there is no way around it.
And you will also need a lot more generator to supply your 3000 T/R modules. And no matter how light your modules is, it will be much heavier than a simple composite or metal airframe

Rafale is planned to be equipped with a 150KW laser weapon also, well before 2020. Switching to GaN will result in weight loss. And the Rafale F3R is getting a new engine upgrade.

GaAs can be built like that too, it called a smart skin, has nothing to do with the material. And you will still have weight, cooling and power problem

It's going to be really fat. GaAs has space and weight constraints.

 

[moon_light]

Sorry, but you haven't understood the OP.

i understand it clearly, it you who keep making up nonsense to support your agenda that rafale is some sort of god like machine

Even the manufacturer website says it.

http://www.dassault-aviation.com/en/defense/rafale/the-sheer-power-of-multisensor-data-fusion/

say what? where exactly does it say Spectra can adapt and develop a jamming profile that suitable to the threat that isn't in its library?

No aircraft does this.

every single one with sensor fusion can do that automatically, even the one without sensor fusion can do that with the help of pilot

They don't. There are only four aircraft known to have sensor fusion, Mirage-2000, F-35, Rafale and F-22.

oh really? , i dont think so
https://www.flightglobal.com/news/a...d-sensor-fusion-to-super-hornet-fleet-376973/

What the other aircraft do is they simply use sensor data which is most reliable to declutter the display. That's not sensor fusion. This psuedo-fusion happens post processing.
On the Rafale, all the sensors send raw data to a single computer. And all that data is crunched to create a single air picture without dismissing data from other sensors. Rafale has pre-processing sensor fusion.

Source? evidence?
The F-35 claim to do exactly the same thing

The most important capability is not even close to ready. The top tester himself says it.

http://www.bloomberg.com/news/artic...till-falls-short-pentagon-s-chief-tester-says

The F-35 is only at IOC, of course it wont have all the capabilities yet, big deal?

The Rafale carries plenty of sensors to avoid that situation. Pulse compression doesn't affect active cancellation. Localization of the signals is 0.1deg.

do you even understand what is pulse compression? it has nothing to do with the ability to geolocating threat so your angular accuracy doesn't matter, point is if you just produce cancellation pulse to every single signal you received no matter if they are in threat library or not, you will ended up jamming ground radar reflection and data link reflection as well as background noise
And where is the source that Spectra have 0.1 deg angular accuracy? what kind of target? ground? air? moving or fixed? etc

Nobody is shrinking the antennas.

So what part do you intended to shrink to get 10 times smaller? the phase shifter? or do you want to shrink the amplifier?

GaN needs lesser cooling than GaAs. Safe operating temperatures are 2.5 times higher. GaN burns at 400degC, GaAs maxes out at less than 200degC.

just because the materials burn at higher temperature doesn't mean you will need less cooling, the heat produced by the equipment still there, if you dont get rid of them, not only your other chips will burn but your fighter will also shine like a Christmas tree in Infrared wavelength

There are 10-20 on the canard leading edge alone, for Spectra. The ones on the wings are much larger. But we need to wait for the layout. The design in the image is generic

so you basically make up the 3000 T/R modules number did you? .

A Mirage-2000 pilot told me the date. There are various news articles also explaining about conformal arrays.

give me a fucking break, it isnt a pilot who told you, i am so feed up with these BS pilot story, the truth is some member on IDF pretend that he is a pilot, then he said a bunch of BS nonsense, and everyone there eat that up

And you need 3000 for the system to be effective. Take the image as an example, the main array has 1000, the two side arrays will have 500 each, that's 2000 right there. The remaining 4 arrays could have 200-300 each, that's 800-1200. Total = 2800-3200.

where do you get all these BS 500, 200 T/R modules number from.? is there any actual source or is this the same BS as last time?

Rafale is planned to be equipped with a 150KW laser weapon also, well before 2020.

No it doesn't, the F-35 is planned to have one, and it has space for one ( replace the fan in the B version) , internal space of Rafale is packed

Switching to GaN will result in weight loss.
It's going to be really fat. GaAs has space and weight constraints.

No it wont, putting 3000 T/R modules along with their cooling and power generator... etc on rafale will make it much much heavier

 

[randomradio]

i understand it clearly, it you who keep making up nonsense to support your agenda that rafale is some sort of god like machine

Why do you think Trappier announced in the French Parliament that not even the F-35 can equal the Rafale?

say what? where exactly does it say Spectra can adapt and develop a jamming profile that suitable to the threat that isn't in its library?

There are French and India professionals on the forum I'm usually on.

every single one with sensor fusion can do that automatically, even the one without sensor fusion can do that with the help of pilot

It can't. Rafale fuses raw data, the others don't. And F-35 is not ready.

oh really? , i dont think so
https://www.flightglobal.com/news/a...d-sensor-fusion-to-super-hornet-fleet-376973/

That's not sensor fusion. That's decluttering processed data.

SH's sensor fusion is completely different. It has separate computers for the radar, IFF, EW etc. After all the data is processed, it is sent to another computer. This computer decides which sensor has managed to extract the most amount of data, and then displays that on the screen. Pilot has to do the rest of the work in order to get more data from other sensors.

What Rafale does is takes raw data for all sensors and then processes it together. It doesn't have separate computers for radar, IFF, EW etc. All the processing is done by one computer by taking raw data from all sensors as input, even MICA's IRST. For example, the computer compares radar signatures of all the returns with the IR signature within the same viewing angle. Then the computer can de-clutter the data, including false alarms, by comparing the two pictures. It can even fuse the best targeting data from multiple sensors, like range from radar and angle from IRST.

Source? evidence?
The F-35 claim to do exactly the same thing

Yes, only the F-35 can do this, along with Rafale, but others cannot.

The F-35 is only at IOC, of course it wont have all the capabilities yet, big deal?

That's among the reasons why Trappier said nothing today matches the Rafale.

do you even understand what is pulse compression? it has nothing to do with the ability to geolocating threat so your angular accuracy doesn't matter, point is if you just produce cancellation pulse to every single signal you received no matter if they are in threat library or not, you will ended up jamming ground radar reflection and data link reflection as well as background noise

No. And I don't think you are able to understand what I'm saying. Geolocating threats is an extremely important requirement for active cancellation. Without that, it is impossible.

And where is the source that Spectra have 0.1 deg angular accuracy? what kind of target? ground? air? moving or fixed? etc

A lot of that is secret.

So what part do you intended to shrink to get 10 times smaller? the phase shifter? or do you want to shrink the amplifier?

Nothing is shrinking. The size is relative to performance. GaN is 10 times better, so you need 10 times lesser number of T/R modules to get the same kind of performance as GaAs. That's why Spectra's current jamming performance has increased by at least 5-10 times using the same aperture size.

just because the materials burn at higher temperature doesn't mean you will need less cooling, the heat produced by the equipment still there, if you dont get rid of them, not only your other chips will burn but your fighter will also shine like a Christmas tree in Infrared wavelength

Your cooling requirement is dependent on the safe operating limits of the HEMTs. You need lesser cooling than you do for GaAs.

so you basically make up the 3000 T/R modules number did you? .

I think you have difficulty understanding what I'm saying. Why don't you read what I originally said.

Go to post 17 and read what I actually said.

give me a fucking break, it isnt a pilot who told you, i am so feed up with these BS pilot story, the truth is some member on IDF pretend that he is a pilot, then he said a bunch of BS nonsense, and everyone there eat that up

Haha.

Anyway, the original date was 2025, open source news from all the way back in 2011. The date was originally posted in Combat Airforces Monthly. But all development has been accelerated because India will also be a funding party. Plus total Rafales to be manufactured has roughly doubled in potential. GaN arrays for Rafale is already ready. If some country wants it, they can have it immediately, within the next 3 years, but at greater expense. The 2021 date is for the French air force.

Rafale will have 360 degree radar coverage from 2021 onwards.

This article from 2011 mentions side arrays.
http://aviationweek.com/awin/plans-further-rafale-upgrades-emerge

Potential radar cross-section (RCS) improvements are under consideration, as are equipping the fighter with additional radar arrays for greater spherical coverage and adding thrust vector control to the two M88 turbojets.

For instance, the side arrays would be used to increase radar coverage.

They also said Spectra will get GaN upgrade by 2020, but that's already been done.
This is the upgraded one.

I would recommend going through this thread first.
https://defence.pk/threads/the-rafale-hidden-beauties-and-its-future.422896/

where do you get all these BS 500, 200 T/R modules number from.? is there any actual source or is this the same BS as last time?

It's logic. You can't have radar performance with lesser than that. The side arrays are half the size of the main array.

No it doesn't, the F-35 is planned to have one, and it has space for one ( replace the fan in the B version) , internal space of Rafale is packed

Rafale's laser comes in wing mounted pods.