Royal Australian Air Force [RAAF] News, Discussions and Updates

[hauritz]

I would concur. I am sure the current version of the Ghost Bat (and for that matter the Ghost Shark) will not be the eventual matured AI technology that is then produced in larger numbers. The AI and the incountry build capability are the crown jewels.

Both are currently limited to surveilance functions, however the Government itself stated we will see an armed AI platform this year.

Ghost Bat/Shark may not be the actual platforms that get weapons, maybe it's a derivation or a new platform that uses the technology explored with these two, as you say Hauritz.

In regards to US involvement, given that AUKUS pillar 2 includes AI, one would view that the outcomes will be shared regardless of US investment or not. So Boeing do possibly get a back door access to the US program, through Ghost Bat if they do it well. I read that there are benefits to an Australian led v US led innovation style project with Ghost Shark (less red tape), so perhaps it helps to keep Ghost Bat as Australian.

Having said that, how much of a stretch would it be to develop an enlarged Ghost Bat with an internal weapons bay for say a couple of LRASMs. It would be a useful option to send a pack of them into harms way to launch a stand off attack against a heavily defended enemy fleet. Or alternatively a stretched shark with torpedos. Perhaps both.

I'm more familiar with Naval systems rather than aircraft, and I'm a bit old school for AI in general, so I'm somewhat of a novice in this area. But I would be interested in others views as to how far off something like this really is. Could we see a gen 2 Ghost Bat with an attack capability this year. I've read that more are being built, however the spec is vague, just that they would be updated versions of the existing ones.

I read the article talking about the USAF's efforts converting an F16 into an AI craft that can aparently do basic dog fighting. If that's the current cutting edge, then an independent AI mini stealth bomber doesn't seem that far fetched.

It seems like just about every man and his dog is building combat drones at the moment. They really aren't that hard to build. Even the electronic brain that goes into them needn't be any more complex than your average mobile phone.

Recently we saw Iran send several hundred drones into Israel. Israel the US and its allies publicly boasted that they successfully shot down 99% of them. I have no doubt that they did. However the reality is that Israel and its allies expended hundreds of millions of dollars worth of very expensive weapons to shoot down what I expect were mostly cheap unarmed drones. We are seeing the same thing on a dailly basis in the Russian Ukraine war with Ukraine exhausting their air defence missiles against relatively cheap drones. We are seeing the same thing from the other side as well with Ukraine launching drone strikes deep inside Russian territory hitting refineries and other infrastructure.

This is why I believe the most important thing isn't the drones themselves. Any stockpile of drones you have will be quickly expended during any shooting war. What is important is your ability to mass produce them. Things like the flatpack cardboard drones might actually turn out to be more effective than more sophisticated and expensive Ghost Bats. Cheap drones will make us have to completely rethink how we will fight wars in the future.

 

[Stampede]

It seems like just about every man and his dog is building combat drones at the moment. They really aren't that hard to build. Even the electronic brain that goes into them needn't be any more complex than your average mobile phone.

Recently we saw Iran send several hundred drones into Israel. Israel the US and its allies publicly boasted that they successfully shot down 99% of them. I have no doubt that they did. However the reality is that Israel and its allies expended hundreds of millions of dollars worth of very expensive weapons to shoot down what I expect were mostly cheap unarmed drones. We are seeing the same thing on a dailly basis in the Russian Ukraine war with Ukraine exhausting their air defence missiles against relatively cheap drones. We are seeing the same thing from the other side as well with Ukraine launching drone strikes deep inside Russian territory hitting refineries and other infrastructure.

This is why I believe the most important thing isn't the drones themselves. Any stockpile of drones you have will be quickly expended during any shooting war. What is important is your ability to mass produce them. Things like the flatpack cardboard drones might actually turn out to be more effective than more sophisticated and expensive Ghost Bats. Cheap drones will make us have to completely rethink how we will fight wars in the future.

This space is evolving rapidly.
Speculative as to how it plays out.
Minor adjustment to the battle space or a complete revolution in military conduct.
Low cost and numbers trumps sophistication.
Land, Sea and Air.
What platforms do we "gamble" upon going forward.
Awkward decisions.

Cheers S

 

[SammyC]

Seems to be two divergent approaches with dones.

• Lots of lower tech and low cost dones to overwhelm a broad defence; or
• Fewer high tech drones to defeat a localised defence.

I see benefits and problems with both. I think there is some analogy (poor as it may be) to the choice of sending in an infantry batalion or an SAS strike team. They serve different purposes.

I do view that drone swarm countermeasures (at least aerial types) are rapidly advancing and cheaper defence systems are emerging (Geopards and Phalanxs are getting a new lease of life, and the likes of Rhinemetal's Skynex are being prioritised in production). So I would view that there becomes more of a balance in time, and simple drone swarm attacks become less effective as defences improve in cost and capability.

More advanced drones with more powerful payloads, that can get in under a defence would therefore still have a role to play. The LRASM for instance is a very smart weapon, but it needs a delivery platform to get it within 400 odd km of the target. This is within the striking and detection range for many defences, putting staffed aircraft in harms way.

Perhaps an approach is to launch something like a half a dozen evolved Ghost Bats (or other similar drone) from RAAF Tindal to cover say the 1,600 and a bit kms to the South China Sea (staying low and within the cover of islands as they approach) and then releasing a couple of LRASMs each when in weapons range of an enemy Naval contingent. Maybe there is a small number F35s to guide them on this journey and provide protection from other aircraft.

The two approaches could work together, with a drone swarm launched nearby to distract attention, while the evolved Ghost Bats get in behind with their payload.

To your point Stampede, the space is evolving rapidly and its speculative how it plays out. But I do think there is a place for both the mass low end and bespoke high end AI platforms.

 

[hauritz]

Seems to be two divergent approaches with dones.

• Lots of lower tech and low cost dones to overwhelm a broad defence; or
• Fewer high tech drones to defeat a localised defence.

I see benefits and problems with both. I think there is some analogy (poor as it may be) to the choice of sending in an infantry batalion or an SAS strike team. They serve different purposes.

I do view that drone swarm countermeasures (at least aerial types) are rapidly advancing and cheaper defence systems are emerging (Geopards and Phalanxs are getting a new lease of life, and the likes of Rhinemetal's Skynex are being prioritised in production). So I would view that there becomes more of a balance in time, and simple drone swarm attacks become less effective as defences improve in cost and capability.

More advanced drones with more powerful payloads, that can get in under a defence would therefore still have a role to play. The LRASM for instance is a very smart weapon, but it needs a delivery platform to get it within 400 odd km of the target. This is within the striking and detection range for many defences, putting staffed aircraft in harms way.

Perhaps an approach is to launch something like a half a dozen evolved Ghost Bats (or other similar drone) from RAAF Tindal to cover say the 1,600 and a bit kms to the South China Sea (staying low and within the cover of islands as they approach) and then releasing a couple of LRASMs each when in weapons range of an enemy Naval contingent. Maybe there is a small number F35s to guide them on this journey and provide protection from other aircraft.

The two approaches could work together, with a drone swarm launched nearby to distract attention, while the evolved Ghost Bats get in behind with their payload.

To your point Stampede, the space is evolving rapidly and its speculative how it plays out. But I do think there is a place for both the mass low end and bespoke high end AI platforms.

Yeah, that is the way I see it playing out. A hi/lo mix of drones with aircraft such as the F-35, F-18, P-8 and E7s acting as network hubs. Actually I would expand that concept over the entire ADF with destroyers, frigates and submarines controlling an array of autonomous vehicles.

I don't know how much research the ADF is doing on finding methods of cheaply eliminating enemy drones but that would be one of my priorities.
Imagine the amount of havoc you could wreak even in peace time by constantly sending cheap, unarmed and unidentifiable drones into another country's airspace. It would cost them millions to intercept and destroy them. Sound like almost the perfect grey zone tactic. This is why drones are so disruptive. You can do stuff you wouldn't usually do because you aren't putting human operators at risk.

 

[south]

@SammyC

Fortunately someone has done a lot of the leg work for you. The Mitchell Institute commissioned a series of wargames where they invited Subject Matter Experts to conduct a series of missions involving Collaborative Combat Aircraft (CCA) in a notional Defence of Taiwan scenario.

Outcomes showed that a mix of CCA increase the mass, survivability and lethality of e fighting force. It highlights included that a mix of CCA, both noderately expensive (on occasion) and low cost/attritable (more commonly) are required. A second insight was that CCA should be survivable enough to reach AIM-120 launch points while equipped with the power and cooling requirements to complete Kill chains.

There’s a lot more at the link; I also highly recommend reading the entire report to anyone who wants to understand some more of the actual challenges around establishing air superiority in East Asia.

 

[SammyC]

Thanks South, interesting and insightful read. The authors didn't pull any punches when articulating why the USAF is in the situation it finds itself. I suspect similar themes could be applied to the ADF, mind you I think the RAAF has done a relatively good job on rejuvenation, better than the RAN at least.

There were a couple of items that caught my attention, one in particular related to the degree of autonomy given to an AI drone. At present most militaries require a human in the loop for kill decisions. The article however, made a point of stating that in highly contested conflicts, comms links can be severed, the exact target may not be known beforehand and response times can be very short. Requiring a human in the link in this environment can be detrimental.

So, giving a drone the ability to select its own target, decide when to fire and even direct another drone to attack, all become combat enhancers.

Perhaps we are closer to a Terminator world than I thought.

 

[hauritz]

Thanks South, interesting and insightful read. The authors didn't pull any punches when articulating why the USAF is in the situation it finds itself. I suspect similar themes could be applied to the ADF, mind you I think the RAAF has done a relatively good job on rejuvenation, better than the RAN at least.

There were a couple of items that caught my attention, one in particular related to the degree of autonomy given to an AI drone. At present most militaries require a human in the loop for kill decisions. The article however, made a point of stating that in highly contested conflicts, comms links can be severed, the exact target may not be known beforehand and response times can be very short. Requiring a human in the link in this environment can be detrimental.

So, giving a drone the ability to select its own target, decide when to fire and even direct another drone to attack, all become combat enhancers.

Perhaps we are closer to a Terminator world than I thought.

As far as I know no missile has a self destruct button or the abiity to return to base. Usually once a weapon is launched towards a target there is no turning back. I can't see why AI controlled vehicles would neccessarily be any different, at least is a full on war situation.

 

[MickB]

As far as I know no missile has a self destruct button or the abiity to return to base. Usually once a weapon is launched towards a target there is no turning back. I can't see why AI controlled vehicles would neccessarily be any different, at least is a full on war situation.

I think he is referring more to who makes the decision (human or AI) on when to fire the missile and at whom, rather than the missile itself.

 

[SammyC]

One of the scenarios mentioned in the report related to sending drones armed with AMRAAMs into an active operational area without a predefined target. It was propositioned that the drone, once launched and in the area could be free to identify hostile targets (enemy aircraft for instance, with no IFF return) and make a decicision to fire on them.

So the human decision is limited to launching the drone and releasing it within an area. The drone has the autonomous decision to engage and fire, without further human approval.

Perhaps it mistakes an enemy aircraft for a passenger liner, but then I'm not sure what the passenger liner was doing in a hot operational area.

Hauritz, its perhaps the next step from launching a missile, whereby the human selected the target and elected to fire. But it is similar in that the human decision ended on launching.

 

[Bob53]

It seems like just about every man and his dog is building combat drones at the moment. They really aren't that hard to build. Even the electronic brain that goes into them needn't be any more complex than your average mobile phone.

Recently we saw Iran send several hundred drones into Israel. Israel the US and its allies publicly boasted that they successfully shot down 99% of them. I have no doubt that they did. However the reality is that Israel and its allies expended hundreds of millions of dollars worth of very expensive weapons to shoot down what I expect were mostly cheap unarmed drones. We are seeing the same thing on a dailly basis in the Russian Ukraine war with Ukraine exhausting their air defence missiles against relatively cheap drones. We are seeing the same thing from the other side as well with Ukraine launching drone strikes deep inside Russian territory hitting refineries and other infrastructure.

This is why I believe the most important thing isn't the drones themselves. Any stockpile of drones you have will be quickly expended during any shooting war. What is important is your ability to mass produce them. Things like the flatpack cardboard drones might actually turn out to be more effective than more sophisticated and expensive Ghost Bats. Cheap drones will make us have to completely rethink how we will fight wars in the future.

I want to see one made of a Carlton draft carton. Drink the box. Unfold the carton. Inside the drone blueprints are there ready to cut out and glue.

 

[Sideline]

I want to see one made of a Carlton draft carton. Drink the box. Unfold the carton. Inside the drone blueprints are there ready to cut out and glue.

As a Queenslander I know that my Bundy Rum carton drone,
would bash the crap out of your Carlton draft drone ;-)

 

[aussienscale]

Interesting to see these 2 heading up the highway this morning, spotted heading north on the Pacific Highway near Byron Bay.

 

[hauritz]

Looks like it could be an F-18, Not sure if it is a classic or one of the newer versions.

 

[Redlands18]

Looks like it could be an F-18, Not sure if it is a classic or one of the newer versions.

Probably classics being taken to a Museum or gate guard somewhere, all the Supers and Growlers are still in service at Amberley. IIRC 7 0r 8 classics were set aside for museums and gate guards.

 

[aussienscale]

Probably classics being taken to a Museum or gate guard somewhere, all the Supers and Growlers are still in service at Amberley. IIRC 7 0r 8 classics were set aside for museums and gate guards.

Agree, would definitely be classics, no wings I could see, but could have been on a separate transporter I did not see.

Either that or they are on their way to Canada

 

[DDG38]

E7 has returned from Op Kudu.
"E-7A Wedgetail A30-006 lands at RAAF Base Williamtown after returning from a deployment at Ramstein Air Base, Germany." Image Source : ADF Image Library

 

[Bob53]

As a Queenslander I know that my Bundy Rum carton drone,
would bash the crap out of your Carlton draft drone ;-)

Yeah but I’d have a better chance of sticking it together after drinking the slab. It’s brain frying stuff….