View Full Version : NightFall - Unmanned Fighters
Nightfall - Machine Autonomy in Air-to-Air Combat (http://www.au.af.mil/au/afri/aspj/digital/pdf/articles/2014-May-Jun/F-Byrnes.pdf?source=GovD)
Although one finds no shortage of professional and academic conversation about remotely piloted aircraft (RPA) and potential unmanned combat aerial vehicles (UCAV), there is a distinct lack of forecasting of their futures on the basis of a tight fusion of tactics, technology, and the enduring truths of air combat. This article claims that a tactically autonomous, machine-piloted aircraft whose design capitalizes on John Boyd’s observe, orient, decide, act (OODA) loop and energy-maneuverability constructs will bring new and unmatched lethality to air-to-air combat. It submits that the machine’s combined advantages applied to the nature of the tasks would make the idea of human-inhabited platforms that challenge it resemble the mismatch depicted in The Charge of the Light Brigade..........
tartare
13th May 2014, 06:44
Very interesting read - thanks for posting.
ShotOne
13th May 2014, 19:54
Interesting read, although I predict most ppruners won't manage all 28 pages!
GreenKnight121
14th May 2014, 05:20
And that unmanned Boyd-fighter will get sniped by a BVR missile from one of those fighters Boyd said was worthless - without said fighter ever being seen by that robot target.
tartare
14th May 2014, 06:52
You think so Green Knight?
I wonder if it'd just snap round in a 20g turn, then lase the missile until it exploded.
ShotOne
14th May 2014, 07:17
Why couldn't a RPA do the same, green knight? While it's a fascinating article I can't help feeling the author hasn't mentioned a key advantage, which is economic. While obsessing with 1v1 combat he misses what shoud be the obvious point that for the price of a manned fighter one could buy twenty RPA's!
Even in real present day scenarios, where very basic Hezbollah drones have been "successfully" engaged by the IDF it was with a Python AAM which cost many times what the drone did.
Roland Pulfrew
14th May 2014, 07:52
An interesting read and does sort of make films like Terminator seem somewhat, and worryingly, prescient. There is a morality question in here which the scientists and engineers seem a little too keen to ignore - do 'we' really want to loose off machines that do all the fighting? Autonomously? With the ability to learn from their mistakes and change tactics sharing that information with other UAVs? Strangely I am just listening to an 'advert' on the radio which was titled "what makes us human?" Where are you John Connor?
I have to disagree with ShotOne's assessment of 20 UAVs for the price of 1 manned fighter though. That is a statistic bandied around much like the "disbanding the Air Force would save millions of £s" often thrown around by green and dark blue personnel, without any real substantiation. Given that our UCAV (note not RPAS) will still require approximately 80-90% of the systems fitted to a modern FJ (the only bits missing are the pilot, cockpit and ejection seat) it is somewhat difficult to justify a 20:1 ratio in price!!
Given that our UCAV (note not RPAS) will still require approximately 80-90% of the systems fitted to a modern FJ (the only bits missing are the pilot, cockpit and ejection seat) it is somewhat difficult to justify a 20:1 ratio in price!! IIRC, for every pilot that goes to war there are another 2 at home - either for roulement or LCR in work-up, so that's a minimum of 3 pilots per airframe. Also assume a reasonable airframe life - since it won't need to support routine day-to-day training flying - so over it's life that might equate to 9-12 pilots, all of whom need training, feeding, paying and a pension. Let's assume a total of £10M per head over a career. That's £90-120M in savings per airframe to offset against the purchase price.
IIRC, then every Lb of primary structure gives a saving of 20Lb in support structure, fuel etc in an airframe. Strip out the pilot and all the support equipment such as seat, OBOGS, cockpit etc, then the airframe can be half the size and much stealthier, or with twice the endurance - so with a CAP at 300nm from base it can half the number of airframes required to support it.
If the basic concept/equipment works, then the savings are certainly there.
So do these machines fly themselves? I thought that there was a stick monkey on the ground piloting remotely; presumably trained and current (cost involved), and with with back up "crew" (cost involved) for constant vigilance. OK, you my save on needing only one operating station (replicating the cockpit), but not if you have multiple UCAVs on station - one driver per one airframe (unless the remote pilot has a brain the size of a planet and is able to control several simultaneosly).
It's not the same as being able to programme a recce drone mission, which requires no routine input from a driver - only monitoring to cater for unforeseen events. A "fighter" UCAV needs to be as reactive as a manned fighter; there is certainly scope for a degree of autonomous action/reaction - but they're only machines linked to computers (fatal error...reboot required...).
Sure, there might be some cost saving, I'll grant that the weight saving on ejector seat and support system removal will improve performance, but does that necessarily mean improved efficiency and effectiveness?
Mister B
An interesting read. Some things to consider as you read it:
The cost of putting a person in the cockpit is only about 4% of the cost of an aircraft and about 6% of the weight.
The author is assuming that the unmanned aircraft would have a full complement of sensors similar to the F-35 so the cost of the aircraft would not be reduced much by removing the pilot. In addition, additional sensors may be needed to account for the visual capability of the pilot.
The cost of software development is a major factor in aircraft development. The cost of software for an unmanned vehicle is much more than for a piloted vehicle to account for system failures normally handled by a pilot (contingency management).
As of now the US Military require the same redundancy in unmanned aircraft as they require in manned aircraft. This is to allow these vehicles to operate over populated areas and in the national airspace. So there is no “savings” from reducing system redundancy.
And for those who want a 20 g aircraft, please understand that as you increase the operational load factor on an aircraft the weight goes up significantly. Not only do you have to account for the obvious wing-fuselage carry through loads, but every component and load carrying element has to be strengthened. For example, the “shelves” which hold the flight control computers have to be strengthened to accommodate the additional load.
There will be a savings in personal costs, but that may be much less than expected, as you continue to have to pay to update the vehicle’s software to take care of the ever changing and dynamic environment of air-to-air combat.
Bevo
You present a sensible and logical view, taking into account the reality of assumed future capability; e.g. the design penalty for increased g loading, and catering for system malfunction and equipment redundanacy.
I had a look at one of our favourite information websites (not always noted for its accuracy) and note that both Preadtor and Reaper ground control stations are manned by two operators - the driver (airframe) and the sensor operator - I guess they share the weapon delivery system mangement (rather than introduce a third person). So any perceived manpower cost saving is not a reality. I'm guessing that somwhere in close proximity is also a supervisor/mission manager or coordinator - something you wouldn't have in a single seat fighter (or with a WSO on board) where the pilot must make rapid autonomous decisions.
I just don't see the use of UCAVs as feasible in the air to air combat scenario as some posters would wish it to be. Great for surveillance with a long loiter capability, coupled with the ability to deliver a variety of weapons and no fatigue penalty for the crew as you simply put a fresh one in the appropriate seat at regular intervals.
Mister B
dubbleyew eight
14th May 2014, 16:42
can you imagine a war where multiple combatant countries launch autonomous drones at each other?
my prediction is that once another aggressor launches drones against the USA the whole subject will be killed by urgent UN Treaty. ...just like the fate of nukes.
the americans might dominate the middle east and surrounds with drone strikes but just wait until they get one of their victim countries pissed enough to launch against the US. The americans will go to pieces just like they did after 9/11.
ShotOne
14th May 2014, 21:56
Bevo, as your post is otherwise persuasive and well thought out, I would like to accept your figure 4% differential for the cost of a RPAS versus a manned aircraft. If, as I suspect though, this refers simply to a manned aircraft minus the components associated with its flight crew, I submit this is erroneous. RPAS aren't just aeroplanes without pilots. They represent a whole new ball-game with its own opportunities and limitations.
No air-defence RPAS has yet entered service so nobody knows what one will cost. I've put forward a figure of 1/20 the cost of a manned aircraft which has been challenged. My figure was a guess/estimate but not a total stab in the dark; the F22 wound up costing $339 million US per unit over its production cycle. A Reaper will cost you $16.9 m. I appreciate they have different roles but do the math, as they say. But even if you entirely disagree, if the cost comes to 96% of a manned aircraft, or anything like, there is no chance whatever of it receiving funding.
smujsmith
14th May 2014, 22:12
Interesting stuff, but, I would hark back to a time when we all had a horror of going to war, based on the real and horrific human cost of doing so. It seems that as years have passed from such wars as Korea and Vietnam, the US (and other) armaments industries have taken on board the rejection by the public of the losses, and headed straight for war by robot. Thus removing casualty figures, the "minimum casualty sales pitch". Lets face it, we will never see a parade in Royal Wootton Bassett, to see the return of a "shot down" computer. In the transition period, we have had the "precision guided, minimum collateral damage" strikes on Iraq, Afghanistan, Libya etc. "surgical" strikes on wedding ceremonies by UPVs across Afghanistan and Pakistan. Most of the British political parties are still in denial at the civilian casualties from the Iraq debacle in 2003. I suspect that Robots are yet another extension of that line of sales pitch. I'm getting on a bit at 61, and may be a bit cynical, but, while I feel for the loss of every crewman in combat, I thank god they are there, using their intellect and humanity to judge whether lobbing that bomb is the correct action. No one in an office in Los Angeles, and certainly no computer driven robot will have that instinct or ability to correctly analyse the situation. There, that's my old fashioned input, a target for the modernists perhaps. War is hard to justify, claiming that war can be prosecuted without casualties on your own side is a rationale to be used by political scoundrels, who have no care for the potential results.
Smudge
finestkind
14th May 2014, 23:12
Somewhat questionable with respect to cost and along the lines of what D/eight indicated. Why do we continually need to upgrade improve etc. our equipment? Cause the other guys got something better.
If it comes to the point whereby the opposition has a wiz bang bit of kit that is far better than your manned piece of kit you have two options. Either get your unmanned bit of kit our rally to have it banned.
Smujsmith.
I agree with you but also need to point out the frailty of the human being. Human emotion will over rule humanity as seen in any conflict. When I have just learned that my family/home has just been taken out do you think that I am going to be overly concerned where I place my load.
GreenKnight121
15th May 2014, 02:06
Why couldn't a RPA do the same, green knight?
Because the article specified that the rpa was to be built according to Boyd's principles - one of which is that BVRs are useless, and fitting provisions for them a waste of airframe volume and space, and thus should never be carried (nor the capability designed in).
Thus the "Boyd-rpa" would not have that capability.
You think so Green Knight?
I wonder if it'd just snap round in a 20g turn, then lase the missile until it exploded.
Sorry, but removing their pilot does not allow for the laws of physics to be suspended.
That kind of G-capability is limited to small airframes such as air-air missiles - an airframe the size needed for the range (fuel)/speed (engines)/payload required for the described mission cannot do that.
Unless the airframe is so heavy that it can barely stagger into the air. Current manned fighters break at 12G or less, and removing the pilot won't allow much improvement.
GK121
Agreed; pretty much what Bevo says as well.
Saving in manpower? Here's an extract from the RAF website (open source http://www.raf.mod.uk/equipment/reaper.cfm):
The Reaper RPAS is operated by Nos 13 & 39 Squadron. The aircraft is operated by a pilot, a sensor operator aided by a non-aircrew Mission Coordinator (MiC). In support of current operations the Reaper RPA is launched from an airfield within Afghanistan by crews deployed in theatre. Once airborne the mission is flown by the crews of 39 Squadron from Creech Air Force Base in the USA or 13 Squadron from RAFWaddington by secure satellite communication before control is handed back to the crew in theatre for landing. The current aircrew of UK Reaper have all been previously qualified as pilots on other military aircraft (such as Tornado, Harrier, Apache, Nimrod, Puma, Merlin and C-130).
So, two crews involved, one consisting of three mission operators (leave aside the groundcrew/maintenace for the moment, as this could be common to both manned and unmanned aircraft) and one that launches and recovers the aircraft.
Don't see much saving there; plus for long missions (14 hours fully loaded) you might have to use two or three replacement crews, especially the sensor operator.
Mister B
Bevo, as your post is otherwise persuasive and well thought out, I would like to accept your figure 4% differential for the cost of a RPAS versus a manned aircraft. If, as I suspect though, this refers simply to a manned aircraft minus the components associated with its flight crew, I submit this is erroneous. RPAS aren't just aeroplanes without pilots. They represent a whole new ball-game with its own opportunities.
Consider the unmanned aircraft from this prespective. Assuming you want an unmanned aircraft with capabilities (range, speed, signature, etc.) similar to an F-35A. Other than removing the equipment for the pilot what systems would you eliminate? The unit cost of most current fighters driven by the subsystems and avionics (about 70% of the cost).
If you want your unmanned vehicle to have the ability to really be in the OODA loop as the article states, you will need at least as many sensors as currently on the F-35 to provide the date needed by the on-board computer.
Assuming a very autonomus vehicle, it may be possible to reduce the controller to a single person to save personnel cost. But I don't see much savings in the vehicle unit cost.
My belief is that the best way to use an unmanned vehicle in air-to-air is as a "wing man" controlled by a manned fighter. The manned figther would have all the onboard sensors and the "wing man" vehicle would be the "shooter", This would allow a much cheaper vehicle, allow the manned fighter to commit a shot, and increase the weapons capability within this mixed flight. For this to work the unmanned vehicle must have a large degree of autonomy so the pilot does not have to spend a lot of time controlling his "wing man".
TomJoad
15th May 2014, 20:56
There is a morality question in here which the scientists and engineers seem a little too keen to ignore - do 'we' really want to loose off machines that do all the fighting?
On the contrary, scientist and engineers are actively involved in the debate Roland. John Connor as you say is indeed alive and well and still battling Sky Net.:ok:
Scientists Debate Killer Robots at U.N. Conference - NBC News (http://www.nbcnews.com/tech/security/scientists-debate-killer-robots-u-n-conference-n103406)
Tom
we are the resistance
ShotOne
16th May 2014, 09:32
Bevo, "what systems would you eliminate". Your question presumes that in designing a RPAS we start with a manned aircraft then remove the pilot bits. That's not the way I see it; it's a new ball game. For starters why would it need to fly at 1000 miles per hour? It's already where it needs to be.
Perhaps I've given the impression I agree with everything in the article. While it was well written and thought-provoking, I feel the authors vision of a hi tech dogfighting UAV constructed of expensive yet-to-be-developed exotic materials is wide of the mark.
Bevo, "what systems would you eliminate". Your question presumes that in designing a RPAS we start with a manned aircraft then remove the pilot bits. That's not the way I see it; it's a new ball game. For starters why would it need to fly at 1000 miles per hour? It's already where it needs to be..
I don't understand your ststement that "It's already where it needs to be".
My point was that except for the system supporting the pilot, all other systems and sensors on current fighters are there to support the air-to-air mission not the pilot. Unless you change the mission requirements you will still need those same systems and sensors to support the basic vehicle flight and provide data to the on-board computer.
ShotOne
17th May 2014, 05:10
Designing an RPAS by starting with a current fighter then removing some systems would make no sense. It would still be eye-wateringly expensive and probably useless.
Perhaps we're looking at different horizons but initially, air-defence RPAS, are likely to follow the example of other RPAS in service. First with appropriate sensors, later weapons. By "it's already there" I meant a key selling point of any RPAS is ability to remain on station for many hours -so it doesn't need to get there at 1,000 mph! So far their development for air defence has lagged other roles. This isn't because it's impossible, rather that we haven't had to fight a war recently where air defence was contested.
Finningley Boy
17th May 2014, 07:59
I suppose if such a system is going to be that good it'll literally fly and think for itself, the term RPAS will in itself be redundant. It would be interesting to see just how close to procuring such a machine anyone is? I know all about Taranis and its mooted Global strike capability, which I suspect is aimed at becoming the true descendant of the Lancaster/V-Bomber/Tornado line. But even here, I can only imagine that such an aerial strike machine will be able to evade/defeat any known manned fighters, air defence missile screens, early warning radar systems and jamming technology, meaning that if this is all on then there is some amazing equipment being devised at the moment.:ooh:
The same would also need to apply to an autonomous unmanned interceptor/dog fighter. It would need to be a literally invulnerable system in respect, certainly, of any possible adversary.
FB:)
Designing an RPAS by starting with a current fighter then removing some systems would make no sense. It would still be eye-wateringly expensive and probably useless.
Perhaps we're looking at different horizons but initially, air-defence RPAS, are likely to follow the example of other RPAS in service. First with appropriate sensors, later weapons. By "it's already there" I meant a key selling point of any RPAS is ability to remain on station for many hours -so it doesn't need to get there at 1,000 mph! So far their development for air defence has lagged other roles. This isn't because it's impossible, rather that we haven't had to fight
a war recently where air defence was contested.
Actually there have been several studies that I am aware of that have looked at the characteristics of an unmanned vehicle for defensive and offensive air-to-air. Unfortunately most of the studies are either classified or proprietary to the companies doing the work. However, I can relate an operations study which centers on the speed of the vehicle. First off, the speed is not needed to get on station, but rather to prosecute the intercept. In looking at the number of aircraft needed to protect a carrier battle group in a 120 degree threat sector on station 200 nm from the CVBG it took eight aircraft with a top speed of Mach 0.9. But with a vehicle with a top speed of Mach 1.5 you only needed four aircraft. The reason is that the higher top speed allowed the aircraft to get into position to successfully intercept threat aircraft from "on station" to anywhere in the threat sector. In addition, thrust needed for speed is also used to keep energy up during a maneuvering fight.
Another characteristic of a lower signature aircraft is swept wings which moves the wing spikes out of the forward quadrant. For example Lockheed's RQ-170 Sentinel has swept wings not for speed but for reduced signature. Another example is a study which looked at taking a B-1 and loading it up with 40 A-A missiles and an F-22 radar. This aircraft could stay on station for seven hours, however, that concept was not found to be adequate because of its lack of maneuverability and speed.
So at least in the studies I am aware of there was not a preconceived idea of what the characteristics of the unmanned vehicle should be.