OT: Revolution in Warfare

[Raellus]

The more content on the wars in Ukraine and the Middle East that I consume, the more I'm coming to think that we are witnessing the beginning of a revolution in warfare as profound as the introduction of steam power, or the machine gun, or motorized aircraft, or radio.

Unmanned vehicles are changing the game, rewriting the rules as we watch on the internet.

Just last night, Ukrainian drone boats sank a Russian corvette with a crew of approximately 45.

A couple of months ago, an off-the-shelf quadcopter costing a couple hundred dollars, armed with a primitive shaped charge that cost even less, knocked out an Israeli Merkava IV MBT fitted with state-of-the-art passive armor and active countermeasures, with a crew of four on board. The total package of that Merkava costs tens of millions of dollars.

There's a profusion of footage coming out of Ukraine of UAVs (mostly FPV types) targeting single soldiers. How cheap has the life of a human soldier, sailor, or airman become? The cost-benefit analyses are getting insane. It's quite alarming.

And we haven't really seen swarm or AI tech in large-scale action yet, but we will soon. See the Loyal Wingman program, or....

https://www.twz.com/sea/interceptor-drone-boat-swarm-vision-laid-out-by-pentagon

Is there a silver lining? Are we drifting towards warfare without human combatants? What would that look like (especially the end game of any armed conflict)?

I'm really interested in what others think about this.

-

[castlebravo92]

The signs were there with the 2020 Nagorno-Karabakh conflict in Azerbaijan. Some of the video that came out of that conflict were pretty sobering. Whole platoons of infantry wiped out by suicide drones at night, captured with a thermal camera on a targeting drone. The attacks looked like cluster bombs, but every single cluster bomb found a human target.

The AGM-65G Maverick costs around $270k per missile (in 1999 dollars, so, like $500k in post-Covid dollars).

$100 will get you a DIY FPV drone, for $5-10 you can get a 1 lb shaped charge or EFP that will punch through the top armor of any AFV in existence today.

A little R&D (that is *easily* within the capability of a non-state actor), you can get semi-autonomous targeting of people or vehicles. If you want to spend a little more, $500 will get you MMWR that will detect people or vehicles, but really, you just need a half-way decent image recognition module capable of running on Android hardware. Much of this tech is essentially off the shelf.

So, for ~$200, you can (potentially) get a semi-autonomous suicide drone capable of targeting and destroying people or equipment or vehicles.

Or, said differently, for the cost of a single Maverick missile, you can get ~1300 of these suicide drones and the shipping container to put them in.

Combine with a few operators, and distribute them around a military base, and in a matter of minutes, you can destroy the operational/maneuver capability of an entire US Army heavy division or take out an entire airbase and all of the aircraft not in hardened hangers.

And while yes we can take out drones - the CBA of taking out a $200 drone with a $1 million air defense missile seems to favor the drone, heavily. Additionally, I don't think anyone has fielded operationally a drone swarm defense. Obviously electronic warfare and jamming is a thing, but jamming works for RPV drones, it won't work on autonomous drones.

Thus the US military (and presumably anyone else worried about this sort of thing) needs to develop or re-develop, muy rapido, an air defense strategy to cost effectively defend against drones. My guess is it will be a combination of HERF weapons and proximity fuzed small caliber (30mm or less) air defense canon.

But yeah, the heavy footprint style of the US military is very vulnerable to drones at least from my casual observation.

[Vespers War]

Ironically, one of the best defenses against drone swarms is likely to be what the US defense industry has been terrible at for decades, which is gun-based air defense. There are medium (35-40mm) autocannon shells now that can have time fuses programmed as they're being fired based on the measured velocity of the round (making it more precise than average muzzle velocity) and the radar range to the target. They're still more expensive than a conventional round, but Rheinmetall was selling 35mm shells of that type for ~$600 each last year, and the price will likely come down (in real terms if not necessarily in nominal terms) as the technology matures. The exchange rate shifts a bit if you're downing drones with a 5-round burst of gunfire at a cost of $3,000 instead of using a million-dollar missile.

The sub-30mm currently aren't seen as being as useful for C-UAS work because the lethal radius of their airburst rounds is much smaller.

The other thing being worked on is electronically frying drones with microwaves, such as the Leonidas (https://en.wikipedia.org/wiki/Epirus_Leonidas) system (there's also Phaser, THOR, and CHIMERA from other manufacturers). Basically, your small cheap drone is going to have pretty much no radiation shielding, because that adds cost and weight. A sufficiently intense blast of microwaves will disrupt the electronics and shut down the drone rather than just disrupt its data link. The capital investment is pretty high, but the cost-per-shot is extremely low, and range is in the realm of 5-10 kilometers. A layered defense with high-energy platforms as the first layer with gun-based air defenses as an inner layer to mop up any survivors would likely limit the effectiveness of the cheap swarm attack, although exactly how effective the defense would be is a fair question.

[castlebravo92]

Ironically, one of the best defenses against drone swarms is likely to be what the US defense industry has been terrible at for decades, which is gun-based air defense. There are medium (35-40mm) autocannon shells now that can have time fuses programmed as they're being fired based on the measured velocity of the round (making it more precise than average muzzle velocity) and the radar range to the target. They're still more expensive than a conventional round, but Rheinmetall was selling 35mm shells of that type for ~$600 each last year, and the price will likely come down (in real terms if not necessarily in nominal terms) as the technology matures. The exchange rate shifts a bit if you're downing drones with a 5-round burst of gunfire at a cost of $3,000 instead of using a million-dollar missile.

The sub-30mm currently aren't seen as being as useful for C-UAS work because the lethal radius of their airburst rounds is much smaller.

The other thing being worked on is electronically frying drones with microwaves, such as the Leonidas (https://en.wikipedia.org/wiki/Epirus_Leonidas) system (there's also Phaser, THOR, and CHIMERA from other manufacturers). Basically, your small cheap drone is going to have pretty much no radiation shielding, because that adds cost and weight. A sufficiently intense blast of microwaves will disrupt the electronics and shut down the drone rather than just disrupt its data link. The capital investment is pretty high, but the cost-per-shot is extremely low, and range is in the realm of 5-10 kilometers. A layered defense with high-energy platforms as the first layer with gun-based air defenses as an inner layer to mop up any survivors would likely limit the effectiveness of the cheap swarm attack, although exactly how effective the defense would be is a fair question.

Yeah, the main problems with HERF (high energy radio frequency) or other EM emitting defenses that I can come up with is:

1) saturation (if I send 100 drones against your microwave defenses, how many get through? That's a $20k attack, even one successful drone may take out a $2 million+ piece of equipment

2) rock-paper-scissors - your EM weapons are going to be strong radiation emitters, which makes them priority targets for HARM missiles or the like (if we are talking about peer competitors).

The reason why you are seeing a lot of drones taking out individual soldiers in Ukraine is likely the same reason why you saw the same thing in the Nagorno-Karabakh conflict. Azerbaijan ran out of higher priority targets to hit, because they had taken out most of the tanks, IFVs, APCs, artillery, and even trucks Armenia had in theater, leaving some poor slob in a trench as the only available target for their loitering drones.

Offensive drones are going to get larger (longer range, larger payloads) AND smaller. A grenade will often have as little as 100 grams of high explosive in it. Still plenty of explosive to wreck sensitive equipment (or propel lethal fragments into a human body). My nightmare is tens of thousands of autonomous micro drones the size of a fist with small explosive charges unleashed on a battlefield to seek and destroy. I really don't know any defense against this except total area denial, and even then, you could deploy an attack from a long range rocket with bomblet dispersal strategies.

[Vespers War]

I think it'll be interesting to see how both 1 & 2 work out. A couple articles I read had one of the Epirus folks describing Leonidas as essentially firing a thousand "bolts" (my term, not theirs) of radiation a second, so a hundred drones may not be enough since they'll need a couple minutes to cross the beaten area of a Leonidas, giving it enough time to fire literally tens of thousands of shots at each drone. And while individual FPV drones can be effective if they hit a vulnerable part of a vehicle or any part of a soldier, individual hits have also been pretty useless if hitting even lightly armored parts of a vehicle, so one successful drone may take out a valuable piece of kit, but it's also possible ten drones will get through and just scratch the paint on some IFVs. Larger drones can have more effect on target, but they're also easier for that second layer of gun-based air defense to engage.

Each shot is also a narrow cone of radiation, so while the tracking radar might be vulnerable to HARM, I'm not sure how well the HARM missiles that rely on continuous lock would do. I'm also not sure how well HARM would survive being blasted by one of the HERF units. And as a third point, most air forces are pretty crap at SEAD - they have few or no dedicated assets for it and most of their pilots don't get enough flight hours to dedicate enough time to SEAD to get good at it. Even with HARM and using long-range rocket artillery to hit low-mobility air defense systems, Ukraine hasn't suppressed Russian air defenses, and Russia's only even partially effective tactic against Ukrainian air defense has been to try to deplete their stock of SAMs with drone and missile attacks. China might be more effective, but the current war has shown that effective SEAD capabilities are part of very few countries' repertoire.

Overall, my feeling is that it'll change warfare, but drones being an unstoppable weapon reminds me of the 1930s claim that the bomber always gets through - true at the time, but a temporary technological advantage that will become untrue in a near-peer environment but might still hold true when one side has an overwhelming advantage either technologically or industrially.

[Raellus]

I read an online article on The Atlantic* last week about this topic and it made the point that, historically, during war, military technology takes great leaps forward due to battlefield exigencies and fear that the enemy will achieve balance-shifting technological breakthroughs first. As a result, those in control of developing and using new military tech make the mistake that the Jurassic Park scientists made- they focus on IF they can't develop deadlier weapons, instead of SHOULD they. The Manhattan Project is a good example.

The various wars going on right now- especially the war in Ukraine- are acting as catalysts for the development of militarized AI.

https://www.twz.com/news-features/drone-warfares-terrifying-ai-enabled-next-step-is-imminent

*Unfortunately, said article is behind a paywall, because it's totally worth sharing.

-

[castlebravo92]

There's been a convergence of, for lack of a better term, dual use technologies that have gotten progressively cheaper and more available.

AI modeling software and AI base models themselves have made it much, much easier and less technically challenging to roll out image recognition software. It used to take years and millions of dollars of R&D to develop image recognition hardware for guided missiles. You can literally use off the shelf models now that run on android level hardware for that, making fire and forget, and fire and seek (without operator control) conceivable for a single person level actor.

Better battery tech + commoditization of drone hardware makes it cheap to deploy drones that have 10-20 mile ranges.

Give a somewhat technically capable dude with $100k, and he could self fund and create a drone swarm that would cripple a modern city.