Missile Defense Computer/AI

[ArkRoyal]

So just doing some reading on weapons/playing DCS or BMS got me to thinking about missile defense in general and something that should have been obvious earlier sprang to mind.

 

Perhaps they do and Ive never seen anything on it, but why dont jets of the modern age have a missile defeat auto pilot? To be clear, this would be something you could turn on or off when you wanted to.

 

When a pilot tries to defeat a BVR missile, he is basically doing some well trained guesswork to defeat said missile. Being show down either comes from entering the missiles no-escape zone on accident or making a bad guess about how to defeat a incoming threat in any other geometry. So essentially, human error. It seems to me like a computer could plot a missile defeat course with essentially nearly 100% chance of success. It has been said on here before that even inside of NEZ or RTR a missile can still be defeated kinetically but that doing so is extremely difficult because the number of successful defeat options is extremely low and hard to fly. Wouldn't a computer be able thread the needle even then? Even if there was only one tiny little loophole to escape, a computer could find it and fly it.

 

Seems like it would work like this:

 

1) aircraft sensors ID thread, or aicraft is programs for theatre threats, or pilot uses info to input this manually when fired on.

 

2) When pilot wants to activate the computer autopilot for BVR. OR it could function as a component of his HSD or hud, simply showing him a flight path he could fly. Perhaps it could show options for both offensive or pure defense routs based on what the pilot wanted to do. IE: it could show the pilot what his options were. How fare he could commit etc. Even once committed what defensive options would succeed.

 

Lastly, for any missiles a nation did not have data for, it seems to me that "close enough" estimates could be programmed into said device, since this would still be more information than the human pilot would have.

 

So has anyone heard of anything like this?

[GGTharos]

No such thing exists. There no perfect that knowledge, even for machines.

[Shmick]

I don't think I'd want my autopilot spontaneously pulling 9 Gs on me, even in a sim :P

[ArkRoyal]

No such thing exists. There no perfect that knowledge, even for machines.

 

Ok so it doesn't exist. I hadn't heard of anything like it either.

 

 

But I dont see any reason why would couldnt or wouldnt. Not perfect maybe, but quite a bit more precise than any human reasoning could produce. Particularly in situations where it might be almost impossible for a human pilot to determine the right course of action due to the number of kinematic solutions being very low.

 

As for being perfect, I also cannot see why not. Presume for the sake of argument that you have at least on missile where you have all the data for it. Physics is absolute, not subjective. It doesn't ever change, and is always the same. If you know the missiles capabilities, and the target planes abilities, then I see no reason that a computer couldn't determine all possible routes of escape right up to the physical limit. Weather would play a role, but this could be input as well, or compensated for by adding in a safety of minus 10 percent or so. This could also be dont for any other unpredictable factors.

 

Essentially, at that point missile defeat is binary. Either there is a available flight profile to live or there is not.

 

This of course bring up another subject as well, what happens when we start using AI for jets completely? The 6th gen jet is supposed to be optionally manned. At some point, you will be able to make a jet that requires no human pilot. It seems to me that at that point air combat both WVR and BVR would become a routine binary exercise. If both AI aircraft know with a reasonable level of accuracy what the characteristics of the other are, then BFM becomes merely a physics experiment. With a human pilot, mistakes get made, estimates of energy and aspect wrong etc. A properly programs AI wouldn't have this issue. For the sake of argument, I am presuming the computers were programmed properly for BFM etc.

 

When two planes merge, (or even well before this really) they would be able to see based on the attributes of the opponent plane all possible BFM solutions. Choose what works. Then one plane wins. Every single time. Perhaps even the other plane realizes it cant win and just runs. Or perhaps it realizes it cant run either. Perhaps nations run sims to find this out before the war even starts and dont even waste their time and just surrender....

Edited January 9, 2016 by ArkRoyal

[NeilWillis]

I think you're flogging a dead horse with this one. There would be howls of derision from the whole community if this was ever implemented, and it'd be banned from all multiplayer servers instantly. Why? because it doesn't exist and is therefore merely a cheat.

 

The pilot would be subjected to violent manoeuvres he couldn't predict, slammed into 9G loads suddenly, and that alone would make the real thing utterly impractical. Add to that the inability of a machine to check it was in clear airspace first. Your wingman would definitely be pretty badly incovenienced if you turned straight into him I imagine?

 

I'd turn my creative genius to designing the perfect coffee machine for the cockpit, or a zero G ashtray for astronauts instead - much more feasible!

[ArkRoyal]

I think you're flogging a dead horse with this one. There would be howls of derision from the whole community if this was ever implemented, and it'd be banned from all multiplayer servers instantly. Why? because it doesn't exist and is therefore merely a cheat.

 

The pilot would be subjected to violent manoeuvres he couldn't predict, slammed into 9G loads suddenly, and that alone would make the real thing utterly impractical. Add to that the inability of a machine to check it was in clear airspace first. Your wingman would definitely be pretty badly incovenienced if you turned straight into him I imagine?

 

I'd turn my creative genius to designing the perfect coffee machine for the cockpit, or a zero G ashtray for astronauts instead - much more feasible!

 

Whoa whoa whoa. Who said anything about doing this in DCS? This isnt in the game section of the forum for a reason....

 

 

Second, all of those things you mention could easily be accounted for.

 

1) computer could warning the pilot before a high G maneuver is sustained. It could also of course be programed to fly withing human limits, just like it is programed to fly within the aircraft limits....

 

2) It can use its sensors to defect friendly planes, even data links could be used for this....There is no reason whatsoever a machine could not detects its own wingman. If you install this system in a plane without the requisite sensors to make it work, thats just like installing a AMRAAM in a plane without a radar.

 

3) If for some reason you installed this on a plane without sufficient sensor fusion or data links to understand its own surroundings, I mentioned that such a system could be either pilot directed and pilot initiated. For example, the ai could just plot courses and the pilot fly them himself. Or if the pilot perceives a problem at any point in the auto verstion...he just takes back control.

[Isegrim]

I did read a while ago that they where testing a system on the Tiffy. Maybe a part of EuroDASS.

 

But it was never a Autopilot. It just displays information about incoming missiles into the HUD/HMS of the Pilot where the best way to defeat those missiles is.

 

Iam going to search for it.

 

EDIT:Uhh its been the German wiki about the EuroDASS so maybe nevermind.

 

https://de.wikipedia.org/wiki/EuroDASS_Praetorian

GER:

Das EuroDASS Praetorian ist das Selbstschutzsystem des Eurofighter Typhoon, welches vom EuroDASS-Konsortium, bestehend aus Airbus Group, BAE Systems, Elettronica und Indra Sistemas gefertigt wird. Die Entwicklungsphase war durch wechselhafte politische Entscheidungen geprägt, was zu Verzögerungen führte. Obwohl die Entwicklung des Selbstschutzsystems, früher als Defensive Aids Subsystem (DASS) bezeichnet, mit über £ 1,5 Mrd. (€ 1,8 Mrd.) die zweitteuerste Komponente des Flugzeuges nach dem Eurojet (£ 1,66 Mrd.)[1] ist, ist am wenigsten darüber bekannt. Als 1987 das erste Konzept erstellt wurde, wurde auf einigen Gebieten Neuland betreten: EloUM/EloGM-Antennen mit aktiver elektronischer Strahlschwenkung, Errechnen einer Feuerleitlösung durch Emitteranpeilung, Millimeterwellenradar zur Raketenortung, Schleppstörsender für ein agiles Kampfflugzeug und die Errechnung von Manöveranweisungen für den Piloten zur Emitteranpeilung und Raketenvermeidung. Durch die Future Offensive Air System (FOAS) Studie dürften weitere Ideen eingeflossen sein, explizite Äußerungen dazu gibt es aber nicht.

 

ENG Google Translator:

The Euro Praetorian is the self-protection system of the Euro Fighter Typhoon, which is manufactured by the Euro Dass consortium consisting of Airbus Group, BAE Systems, Elettronica and Indra Sistemas. The development phase was marked by changeable political decisions, which resulted in delays. Although the development of self-protection system, formerly known as Defensive Aids Subsystem (DASS), designated by about £ 1.5 billion (€ 1.8 billion), the second most expensive component of the aircraft after the Eurojet (£ 1.66 billion.) [1 ] is, is the least known about. As in 1987, the first concept was created, has been entered in some fields Neuland: ESM / EloGM antennas with active electronic scanning, calculating a firing control Emitterlocating, millimeter-wave radar for missile tracking, trailing jammers for an agile fighter aircraft and the calculation of maneuvering instructions to the pilot to Emitterlocating and missile avoidance. By Future Offensive Air System (FOAS) study should be incorporated more ideas, explicit expressions, there are not.

Edited January 9, 2016 by Isegrim

[Bushmanni]

I have seen a Finnish study about algorithm for a flight director system that guides you to steer an optimal missile avoidance maneuver updating the calculations in real time. The study didn't consider if it's possible to know the location of the missile (required for the calculations) or have enough calculation power available to do the calculations in real time though. Tracking the incoming missile while maneuvering is hard and it's also hard to make a system that can estimate the missile position using educated quess better than human.

[Exorcet]

One possibility for not using this is that it might not be as good a tactician as a pilot. If you have multiple options for avoiding a missile, which do you choose? Do you want to extend afterward, or do you want to try to position yourself for a return shot?

 

Also the pilot should be more concerned with avoiding being shot at all. Having a missile come after you is a bad situation.

[GGTharos]

1) computer could warning the pilot before a high G maneuver is sustained. It could also of course be programed to fly withing human limits, just like it is programed to fly within the aircraft limits....

 

What does it do when it knows it can't out-maneuver the missile? Eject?

 

There is no reason whatsoever a machine could not detects its own wingman.

 

There are plenty of reasons for not being able to IFF another aircraft.

 

3) If for some reason you installed this on a plane without sufficient sensor fusion or data links to understand its own surroundings, I mentioned that such a system could be either pilot directed and pilot initiated. For example, the ai could just plot courses and the pilot fly them himself. Or if the pilot perceives a problem at any point in the auto verstion...he just takes back control.

 

What does the system do if it can't get missile distance or velocity? (It won't)

[ArkRoyal]

What does it do when it knows it can't out-maneuver the missile? Eject?

 

 

 

There are plenty of reasons for not being able to IFF another aircraft.

 

 

 

What does the system do if it can't get missile distance or velocity? (It won't)

 

 

 

1.) Your point exactly? At this point the aircraft is dead. Getting a cue from the aircraft computer to tell the pilot he can either take his chances with counter measures or eject is alot better than crossing your fingers when your dead and dont even know it.

 

2)Once again....so? A data link could easily keep the planes in contact is nearly every situation. Add in a system like the F-35's DASS, and your even better off (which can also IFF). Not to mention IFF here isnt the point. The pilot would have control over the system, so he could choose to disengage it if he saw a collision coming. If he didnt see it coming, we'd still be getting a collision missile defense system or not.

 

3) I was never implying you had that information in first place. Only that you know you have been launched on. Once you have been fired on, the computer could take into account the launch platforms kinematics and determine what it is and is not possible for the missile to do. Missile type could either be selected by the pilot or computer depending on how you want to rig it. Ideally, it would be designed to allow the pilot to choose the threat after feeding possible threats to the pilot. IE: hostile platform was Su-27SM, range/speed/aspect etc was blank blank blank, pilot chooses threat from possible missile selection of enemy plane, computer offers defeat options based on pilot selection and other factors. Essentially, the computer shows the pilot profiles the the enemy missile simply couldnt fly based on the launch conditions. Pilot selects one based on his intentions (for example, does he want to go completely defensive or not.) In fact, such data could inform the pilot when he absolutely needs to go full defensive, by showing him the point of no return.

And that is going off my original assumption that no further information is received on the missiles status. Some modern sensors my very well be able to track the missile at certain ranges.

 

 

Not that my ideas matter, because the system already exists.

 

And appears to function almost exactly I envisioned myself,

 

". By identifying the opponent's platform based on emitter identity or suspected target type, based on systems that are in the inventory of the enemy, the potential weapons load of the opponent is determined, as well as their effective range and tactical ability. [36] These libraries are for the Euro Fighter operators freely programmable and the current threat level can be adjusted at any time. [32]

 

When flying with high g-loads information is sent from the Flight Control System (FCS) to the ESM to take into account the bending of the wings in the position determination of the targets. The ESM appreciates the distance to the destination, based on the signal amplitude. The DF accuracy is higher with less than 1 ° as the CAPTOR radar. [32] [37] Due to its high angular accuracy, the system can also be used for geolocation of emitters and fire control. [38] [13] [39] [40] [Ah. 1] The positioning of air targets is challenging since they move in an unknown distance to an unknown price and unknown speed. In order to solve the problem two Kalman filter for a recursive Interacting Multiple Model (IMM) is used, which the antenna positions Interferometriemessungen for angle determination, the Pulse Descriptor Word (PDW), mission data, real-time constraints and track data used by other sensors distance, speed and heading of the target output. . If the angle change is too low, for example, because the transmitter is at a great distance or in front of the aircraft, there are two ways: [36] [Ah. 2]

 

The wingmen send over the data link angle measurements and the PDW. This data is angular, signal-to-noise ratio, SEI (Specific emitter ID of the threat / target radar) over time correlates. The own machine and the wingmen can so determine the target position, because the distance between the own aircraft is known (triangulation.) [36] [41] do not have this option available, accesses the second option: The DAC calculates whether through the emitter there is sufficient information in order to give the pilot a maneuver instruction on the HUD. If this is the case, there are two standard maneuvers are available: the 2-turn two 90 ° curves with a straight line to be flown in between, so that the direction of flight after the maneuver with the foregoing is the same. The second maneuver is sinusoid, in principle, the emitter is served here in a zigzag. The overlays in the HUD are like while dodging missiles; an arrow indicates the direction and g-force to, in addition to the price and the duration of the maneuver are displayed. [36] [Ah. 2]

 

In the event that no antenna in vertical tail exists (see below), there is a second patent by BAE Systems, which can calculate the target positions in three dimensions with azimuth measurements maneuvers overlays and data links. The data processing is more complex here with 12 Kalman Filter as angular measurements, the PDW and other things must also be correlated internally in a database. The Interacting Multiple Model differs here, whether the target air maneuvers or maintains its course. Otherwise, the data processing is identical to that described above. The maximum range of the method is specified with 120 nm (216 km). [42] [Ah. 2]

 

Due to the principle this is only possible in the front hemisphere, since the right rear pod contains the towing jammers. In the rear hemisphere is only a rough indication of the angle (6-18 GHz or 32-38 GHz) or sector (0.1-6 GHz) as the bottom left, top right, etc. possible.

 

Electronic Warfare [Edit]"