How Hot is Ukraine Gonna Get?

[kimbosbread]

It just occurred to me that one awesome advantage of smart spider mines is they could have IFF, so friendly vehicles and soldiers could operate with minimal danger in their own minefield, while enemies would have a rough time of it. Presumably if this would be possible, they could also recognize civilians or animals and ignore them.

[poesel]

15 hours ago, Battlefront.com said:

Actually, I don't presume the attackers are networked.  In fact, I presume that they are fully autonomous.

You wrote about a coordinated attack. For coordination, you either need a third party observer, communication between the drones or passive observation between drones.

In the first case, the observer is at a distance and needs to communicate with the drones. That makes himself a high priority target, and the distance diminishes the value of the observation.

In the second case, the drones need to emit something (radio, IR, acoustic) which makes them detectable in principle. But I guess whoever designs this will make it low noise enough, so that it can be used for detection but not targeting.

The third case is hard because the drone can only use passive sensors to detect others in the same mission. Since the drones are designed to be hard to detect - well... I guess possible, but a few years down the road. That would be real ninja drone swarms.

 

15 hours ago, Battlefront.com said:

But networking doesn't solve the problems of friendly fire coming from point defense weapons.  See previous post.

If I had a bunch of networked vehicles, then each would know where everyone else is at any time and not shoot in their direction.

We are talking about small calibers here. So even if I had to shoot my shotgun equivalent in the direction of something Bradley-ish I would do it. Anything I could do to the Brad is less worse than what the drone does.
And if theres my own infantry directly behind the drone coming at me I would still shoot. That is one CAS vs 3+ in the vehicle. War is no Ponyhof.

 

15 hours ago, Battlefront.com said:

Lastly, the defense needs to be a very heavy emitter. That opens up all kinds of lethal possibilities for the enemy to exploit.

No, he doesn't. Everything I named so far is passive. Being a heavy emitter makes targeting easier. So PD may need to switch on its radar when the attack has been detected and is imminent. But then you are not giving much away since the enemy already knows you are there.

 

15 hours ago, Battlefront.com said:

Well, I tried to give you some credit, but if you insist on being 100% wrong then who am I to argue

I would like to ask you to stop telling other people that they are just wrong. I perceive this as rude, even if you put a smiley to it.

 

15 hours ago, Battlefront.com said:

The only way to get a bullet reliably onto a fast moving target is to be able to predict it's path.  Once the path is unpredictable, the chance of a miss are almost assured when talking about the sort of speed and maneuverability of a drone.  Airburst munitions mitigate that somewhat because "close enough is good enough", but the systems that can do that sort of thing are very expensive and big.  Two things that mean they aren't a practical solution for anything other than a limited number of fixed installations.

You did not follow my example calculations. Depending on the speed of the bullet and the amount of g's the target can pull, there is a distance where the target just doesn't have enough time to move away enough from the current, predictable path.
Enlarging the kill area with a shrapnel weapon (like a shotgun shell) makes hit probability even larger.

Airburst needs at least a 20mm shell IIRC and that is IMHO too big for this application.

 

[hcrof]

1 hour ago, Battlefront.com said:

I got that link to load after disabling my ad blocker.

Yes, that read like any other trade show summary.  Lots of slick sales pitches for systems that hope to find customers, but in the end few will.  People familiar with such shows may even note there are some that come every year and yet nobody is buying.

What I saw in this was promises by companies that if they are given enough money they can come up with at least partial solutions.  Adapting Kongsberg RWS is probably the most practical of everything I saw in there, but again... it has serious limitations which we've already discussed at length.  One that I did not see mentioned is the inability for these weapons to load C-UAS munitions on-the-fly.  That means having a dedicated C-UAS vehicle, which gets us back to the primary weakness of point defense which is that it can only defend a point.  No ability to defend a column from lateral attack in complex terrain, for example.

Steve

I don't believe that a MG is any good for shooting down drones if they are manoeuvring to avoid being hit.  See my previous post on the subject:

The best they can do is force them to manoeuvre which might make them hit armour instead of a weak spot. Canister or airburst from an autocannon is the only thing that will reliably do it at range, although as previously mentioned some kind of shock wave/blast would be effective at point blank if you knew when to pull the trigger. Otherwise, if you know a drone is incoming, smoke or dazzling would be far more effective.

The key in every case is to know you are under attack before the drones start to hit you. 

Edited July 7 by hcrof

[The_Capt]

8 minutes ago, poesel said:

No, he doesn't. Everything I named so far is passive. Being a heavy emitter makes targeting easier. So PD may need to switch on its radar when the attack has been detected and is imminent. But then you are not giving much away since the enemy already knows you are there.

Isn’t the problem with all passive detection stand off range?  Drones are small, made of plastic and can fly between trees.  So if a vehicle platoon, all networked and covering arcs like a B17 flight, is only using passive detection wouldn’t that allow the drones to get much closer than if they were using active? Clutter and obstacles are going to be murder on passive detection of a two foot drone coming in a very low altitude. In fact, terrain specific would this not result in ambush zones all over the place as the vehicle detect too late and get overwhelmed?  Switching on “after an attack is detected and imminent” could easily been the difference between being dead or alive.

On the other hand, passive tech is definitely going to make headway and with AI/ML separating signal from noise will get better.  Sounds like a competitive environment.

[ASL Veteran]

 

[poesel]

Please note that I'm not arguing against this post and the conclusion of it. Just some details.

4 hours ago, The_Capt said:

So effective fully autonomous PD is going to use the same technology as those fully autonomous UAS.  In fact the technology to stop all those small fast moving drones will need to be more robust and complex by virtue of the difference in target profile.  C-UAS is shooting at a small bird sized target, those UAS are shooting at a large metal vehicle.

So what?  The asymmetry in target profile creates asymmetry in targeting technological requirements.  One cannot say that the defensive technology is right around the corner yet offensive autonomous UAS/UGV are over the horizon…because they have shared technologies.  What is fundamentally different is the targets and those target profiles.  So if we invent a super guns system that can detect and effectively engage a drone swarm, we invent the same technologies to make those drone swarms more effective…and very possibly cheaper.

I can speak only for me, and I don't think defensive technology (as in counter drone PD) is right around the corner. I'm just saying it is possible. We may see it in this war depending on how long it lasts or in the next. When this technology comes, the drones will have to adapt.

We have seen the beginning of autonomous drones. This is a logical reaction to increasing EW. The first step in the arms race for this rather new type of weapon. Other steps will follow.

 

 

4 hours ago, The_Capt said:

The problem this entire debate is that one side has an open loop - things are changing and we do not know how to adapt yet. The other is a closed loop. - things are changing but we do know how to adapt…make things like they were before.  So defensive technology looks promising, while offensive technology is “far off and risky”.  This is inconsistent and frankly makes no sense as both offensive and defensive share the same technological base.

I haven't read everything, but that would be a really strange argument to make. Offensive drones already exist in the thousands.

I haven't seen anything against kamikaze drones except infantry shooting their rifles.

4 hours ago, The_Capt said:

One cannot choose which technology is going to have an impact, they all are.  The problem we are facing now is not one of asymmetric technology, it is of asymmetric vulnerability.  A small drone is small and low weight, so it is fast and can hide in environmental clutter.  It uses a lithium ion battery (most of them) so is relatively cool. It makes noise but low energy noise compared to background.

The battery will get merely warm. However, the motors are very hot. Quadcopters push a lot of amps through small engines.

There are designs for noise reduced propeller blades. Still, this will never be noiseless. And you emit in a certain frequency range you cannot change arbitrarily else your copter falls out of the sky. Good microphone + good filter and you can detect a quadcopter. I guess if something would exist that can shoot drones down, we would already see a microphone network in Ukraine.

 

I agree with the rest of the post and the conclusion.

[ArmouredTopHat]

8 minutes ago, The_Capt said:

Isn’t the problem with all passive detection stand off range?  Drones are small, made of plastic and can fly between trees.  So if a vehicle platoon, all networked and covering arcs like a B17 flight, is only using passive detection wouldn’t that allow the drones to get much closer than if they were using active? Clutter and obstacles are going to be murder on passive detection of a two foot drone coming in a very low altitude. In fact, terrain specific would this not result in ambush zones all over the place as the vehicle detect too late and get overwhelmed?  Switching on “after an attack is detected and imminent” could easily been the difference between being dead or alive.

On the other hand, passive tech is definitely going to make headway and with AI/ML separating signal from noise will get better.  Sounds like a competitive environment.

I think this is why Poesel and I are suggesting any PD system is going to be relatively short range and probably firing a shotgun type shrapnel system instead of a proper gun. Researching the subject seems to indicate militaries are focussing on the 30mm airburst, so perhaps non explosive bullets are not the way to go. Though this seems to be for medium range engagements?

The idea is not that this would be a silver bullet, but something to pick off a few drones that make it through. This doesn't need to be a big fancy system but instead really should be as small and relatively simple as possible. Again layered into numerous systems / drones that are doing the defending. A shotgun type weapon could be happily fired into friendly vehicles without too much concern as well, which limits the friendly fire or collateral. 

As you say, passive sensors will have the ability to do this in the future, and if it complicates a swarm drone kill chain it might just be worth it. Of course it wont be fallible and different approaches can and will overwhelm such a system potentially. Again the notion is its better than nothing. Militaries are clearly set on trying to preserve manoeuvre. 

[poesel]

16 minutes ago, The_Capt said:

Isn’t the problem with all passive detection stand off range?  Drones are small, made of plastic and can fly between trees.  So if a vehicle platoon, all networked and covering arcs like a B17 flight, is only using passive detection wouldn’t that allow the drones to get much closer than if they were using active? Clutter and obstacles are going to be murder on passive detection of a two foot drone coming in a very low altitude. In fact, terrain specific would this not result in ambush zones all over the place as the vehicle detect too late and get overwhelmed?  Switching on “after an attack is detected and imminent” could easily been the difference between being dead or alive.

On the other hand, passive tech is definitely going to make headway and with AI/ML separating signal from noise will get better.  Sounds like a competitive environment.

Yes - sitting in a forest is probably the worst if you are attacked by a drone swarm (even if you had a defending swarm of your own). OTOH a desert is not drone country...

If we take Ukraine as test field, I guess it would be okish for the defender. Large open space between tree plants have not much cover for drones. But terrain is a bitch - we all know that.

A defender will need to think different about what angles of attack he needs to cover against drones in the future.

[ASL Veteran]

 

[ArmouredTopHat]

3 hours ago, hcrof said:

Ok, FPV drones. Here is my current take, apologies for the monster post but its a big subject, even ignoring smart artillery, NLOS ATGMs etc.

Lets start by laying out what I mean by an FPV drone:

Current FPV drones: $1000, 3 man crew (operator, navigator, technician), carries up to a 2kg PG7 warhead which can penetrate 250-500mm RHA. Analogue video feed and digital control signal. Artisanal. Unsafe. 5-10km range. Cant fly in bad weather or at night. Needs to be guided in by an ISR drone or sweep a known location like a road. 

Future FPV drones: $3-5000?, 2 man crew per swarm (commander, technician), On-board terminal guidance and swarming, automatic (pre-planned) route following. Similar warhead performance to now but with safety switch so drone can be returned. Assume more aerodynamic but efficiency cancelled out by more computing power requirements. 5-10km range. Cant fly in bad weather. Mass produced but with modular warhead, optics and control system. 

Note that I looked into EFP warheads The_Capt and unless you know something I don't, they don't seem realistic for this application. SMArt and BONUS seem to use 12kg(!) submunitions with worse penetration values than a 2kg PG7 warhead. 

So lets look at a scary drone swarm scenario: an AWACS aircraft picks up a group of moving vehicles 10km behind the lines and an orlan-type drone confirms it is a mixed battlegroup of tanks and IFVs. A swarm of 50 future FPV drones is launched from a lightweight trailer and dispatched to intercept. They navigate there using INS, with occasional updates from the controller on the actual target position. They travel above treetop height but when the attack is confirmed they split into 2 groups and most drop to 1-2m above the ground, with a few going high for situational awareness to coordinate the attack. The 2 groups attack from 2 different directions, timed to arrive at the same time. Since this is beyond the maximum return distance and therefore all drones will be expended anyway, groups of 3 attack each vehicle simultaneously, jinking to avoid fire and targeting tracks and other weak points. One of the 3 drones carries a claymore charge instead of a shaped charge to damage optics and radars from a distance before the 2 AP drones attack. Up to 16 vehicles can be disabled in this one attack which renders the whole battlegroup ineffective using about 250kg of munitions, costing maybe $250,000, with the logistics burden of moving a single trailer into position near the front lines. Artillery is then used to destroy the immobile damaged vehicles, generously say 5 shells per vehicle for 80 total (4000kg of munitions costing $320,000 from a battery of 40 tonne SPH). 

So lets use the defensive onion to try and stop this attack:

Don't be seen: Focus on deception and EW: target enemy orlan-type drones with roadrunner or FPV drones to avoid identification so whole swarms are wasted on decoys and spoofs. Effects need to be massed not platforms: move in small groups carrying long ranged weapons. Use indirect firing weapons. EW: jam video feed from ISR drones, jam controls to prevent them re-tasking (assume they are on autopilot for robustness), jam radar. Target AWACS with missiles and ground based radars with artillery. 

Don't be acquired: Smaller platforms using multispectral camouflage can use more terrain as cover. Fast movement to increase search times for drones and force them to repeatedly get updates from the controller (which can be jammed/targeted). Use cheap pickets to detect FPV drones from a useful distance possibly using acoustics or scanning for control signals. 

Don't be hit: Pop smoke upon acoustic detection of FPV-type drones. Shoot down incoming drones with APS, self defence using autocannons (pointed in the right direction via acoustics if needed) or defensive drones (note that defensive drones will always be lighter/cheaper/faster than offensive drones since they carry a smaller battery and payload). Use your own drones to find and hit the enemy drone launch platform. Smoke+radar+40mm canister shot from autocannons. Directed overpressure wave (think RPG backblast) to knock drones down at close range.

Don't be penetrated: Swap frontal protection for all-around protection and reduced weak spots. Reduce power requirements (weight) to make it easier to protect weak points like engine ventilation. Laser dazzlers to degrade FPV accuracy so they hit armour not weak spots. Reduce crew to reduce internal volume and increase average armour effectiveness. 

Don't be killed: Redundancy so the platform can still move and ideally operate when damaged by small FPV warheads (things like tracks are a weak point so fix that). Fewer humans on board. Spall liners, fire suppression, suspended seats. 

Maybe the onion above doesn't work at all, maybe it is super effective and all drones are eliminated, or maybe 1-2 vehicles are disabled and the mission continues. I don't think it is super clear cut either way and a lot of testing is required to work that out. Also this is just one scenario: maybe future FPVs are super effective against fixed defences like trenches as well as vehicles? What does that even mean for future warfare? 

In conclusion the situation is unknowable right now since we are still in 1916 and trying to figure this all out. While I lean towards drones having the advantage (especially when used in combinations with other systems to introduce dilemmas), there are lots of clever things you can do to blunt that edge and it also doesn't necessarily mean the death of manoeuvre. Remember too that the tank was only one tool of many that ended the stalemate in WW1; we will likely need tools other than drones in 2050. 

One thing I do know though, is that a) vehicles and tactics will have to change radically and b) the drone/counter drone battle will be likely be a whole new branch of warfare going forward. 

Excellent post

[ArmouredTopHat]

55 minutes ago, kimbosbread said:

It just occurred to me that one awesome advantage of smart spider mines is they could have IFF, so friendly vehicles and soldiers could operate with minimal danger in their own minefield, while enemies would have a rough time of it. Presumably if this would be possible, they could also recognize civilians or animals and ignore them.

Such autonomous weapons have the capability of being very humane....or very inhumane. The thought of such smart munitions being programmed to specifically go for say children out of a potential target group as a weapon of terror is sickening to say the least.

It does make me wonder if the worlds powers will actually rule on limitations to such weaponry much in the same way as NBCs. 

[ASL Veteran]

 

[Blazing 88's]

3 hours ago, Probus said:

Or maybe BFC will say screw it and stick to expanding their WWII sims.

!!!!!!!!!

[Probus]

6 minutes ago, ArmouredTopHat said:

It does make me wonder if the worlds powers will actually rule on limitations to such weaponry much in the same way as NBCs. 

I wondered the same.  I really hope so.  But how do you enforce it if the country using them can build them organically?

[chris talpas]

Perun is back this time discussing recent escalations by Russia and what has been or could be the allied response.  As always both informative and entertaining.

 

 

[ArmouredTopHat]

Already seeing a growing emphasis on cheap detection towards drones now. The focus on hunting those much more expensive reconnaissance drones seems especially significant, Ukraine would benefit a lot from constraining that recon capability, if only to deprive Russia of targeting in the rear areas. 

Edited July 7 by ArmouredTopHat

[Butschi]

15 hours ago, chrisl said:

Yes, I'm well aware of the whole context thing.  That's really the only way they could predict other words, isn't it?

Technically, I was talking about the attention mechanism in transformer models. You can use a lot of other architectures in order to predict a word as a function of a bunch of other words that don't have this feature. They also have worse performance, of course.

15 hours ago, chrisl said:

And there really isn't any reasoning behind the answers.

Excuse me but that sounds like a really big claim to make, given that a) there a papers out there that research reasoning capabilities in depth and come to the conclusion that LLMs have at least basic capabilities, b) many in this field use LLMs precisely because of their reasoning capabilities and c) you can actually ask e.g. ChatGPT to reason how it arrived at an answer. Can I kindly ask you to back up your claim?

Sure, LLMs are not AGI and their answers and reasoning are often flawed. That doesn't imply they can't do any reasoning at all, though.

[Blazing 88's]

37 minutes ago, ArmouredTopHat said:

...It does make me wonder if the worlds powers will actually rule on limitations to such weaponry much in the same way as NBCs. 

This is where it will be heading towards imo. Or maybe only to be used defensively.

[Blazing 88's]

Thread relevant:

[cesmonkey]

The new British Defence Minister is already in Ukraine meeting with Zelensky:

 

[TheVulture]

17 minutes ago, cesmonkey said:

The new British Defence Minister is already in Ukraine meeting with Zelensky:

 

Second-hand quoting of a journalist on what the UK has announced as additional support today:

Upon taking office the Defence Secretary immediately asked for extra support to be provided to Ukraine which was readily available and meets their needs for the battlefield against Russia. This new package includes:

• A quarter of a million of 50 calibre ammunition

• 90 anti-armour Brimstone missiles

• 50 small military boats to support river and coastal operations

• 40 de-mining vehicles

• 10 AS-90 artillery guns

• 61 bulldozers to help build defensive positions

• Support for previously gifted AS-90s, including 32 new barrels and critical spares which will help Ukraine fire another 60,000 155mm rounds

John Healey also directed officials to ensure that the promised package in April of military aid is accelerated and delivered in full to Ukraine within the next 100 days.

Not seen any official confirmation of details yet.

Although this is a new Labour government replacing the previous Conservative government, so its always possible there are some political games going on of announcing as new stuff that was already promised but not yet delivered. But at the least the new government is signalling that it is going to continue, and possibly increase, UK support for Ukraine.

 

(Note for confused Americans: Brits use 'government' in the same way that Americans use 'administration' - so read that as 'administration' in the above)

[kimbosbread]

2 hours ago, poesel said:

You wrote about a coordinated attack. For coordination, you either need a third party observer, communication between the drones or passive observation between drones.

In the first case, the observer is at a distance and needs to communicate with the drones. That makes himself a high priority target, and the distance diminishes the value of the observation.

In the second case, the drones need to emit something (radio, IR, acoustic) which makes them detectable in principle. But I guess whoever designs this will make it low noise enough, so that it can be used for detection but not targeting.

The third case is hard because the drone can only use passive sensors to detect others in the same mission. Since the drones are designed to be hard to detect - well... I guess possible, but a few years down the road. That would be real ninja drone swarms.

Nope, that’s all wrong. I think there are 3 likely scenarios, from more distributed to less:

1. Drones don’t communicate, and just attack based on internal assumptions: Roll a dice on which vehicle to attack, modulo priority (fuel truck, comms, etc.), and roll a dice on how long to wait. You lose coordination, but there’s no emissions.
2. Gossip protocol, where a few drones will send a short message, and this will be guaranteed to reach a certain amount of drones with a minimum of communication. You rely on statistics to say target X will get hit with probability Y.
3. Central coordination, where a few drones are elected supervisors, and coordinate which drones to send to which target and when.

This is completely pedestrian stuff; software engineers like me who build systems running across hundreds of thousands of servers for your FAANG companies have been designing and implementing algorithms for minimizing communication and maximizing coordination for a decade or two (and HPC people before us).

And even if you have central coordination, without humans in the loop the loop is going to be very fast- milliseconds for decisions to propagate across a swarm. Your defensive systems won’t be able to react to this unless they are communicating and being noisy!

[chrisl]

3 hours ago, paxromana said:

I remember reading about German RDF prroblems in Crete ... the Resistance was tired of having German RDF vans pick up their signals and home in on the Radio Sets. So they ran wires from one set through to three aerials at the points of a triangle.

The German RDF vans would get the signal as coming from the centre of the triangle ... which is where the Radio Set was not.

I presume that this specific trick may not work against modern EW ... but running wire from the actual controller to a remote broadcast antenna would be a possible solution. Or you could have direct LOS low power microwave band or laser transmitters set up which, as I understand it, are virtually undetectable unless you happen to luck into the direct LOS of the transmitter-receiver link. 

I was a Company Sig in my Uni days and Signallers ran a lot of landline wire routinely. And that was in an Infantry Regiment (Battalion to non Commonwealth types) ... add some wire and some signallers to a Drone detachment and you have at least a potential solution?

Or is this too obvious?

Fiber from the operator to any broadcast transmitter/receiver (and you can bounce around who's transmitting among several).  Optical or microwave to an airborne relay, optical to a relay at the swarm, and only then omni-directional RF to swarm members.  It's the EM equivalent of hiding behind 7 proxies.  And the swarm members will be mostly autonomous and get "mission" commands infrequently rather than realtime joystick control.

[ArmouredTopHat]

I really do wonder how long this can be tolerated for until something suddenly snaps.

[ArmouredTopHat]

Far right in france did quite a bit worse than expected, which is a small relief. 

Edited July 7 by ArmouredTopHat