Hacker News Reader

Comment by jimt1234 | Hacker News Reader

Does anyone know anything about current defenses against large-scale drone attacks? The scenario I'm thinking of is 100+ small drones, each with an explosive device attached (maybe a grenade), all programmed to attack a specific target, at a specific date/time. The defenses that I know about (shotguns and nets?) might stop, maybe, half of the drones, but that still leaves 50 grenades, all set to explode on the same target. I just don't know how a target would defend against an attack like that - GPS jamming?

View on HN

If you want to know more about the state of the art here, read more about the Ukraine war, preferably on Slavic language sites. It seems the main defense is electronic warfare, either jamming the GPS or the control signals. This is leading to an autonomy arms race at the moment, as a drone with a preprogrammed mission and optical flow sensors is much less vulnerable to jamming.

The future of MicroAI is going to be interesting for military purposes. Just think about the AI capabilities of your cell phone these days, then thinking just how a little bit more intelligence on a dumb weapon could make it far more deadly.

Yeah will be cool once “hallucinates” and kills friendlies. Depends which side uses them i suppose.

https://defensescoop.com/2023/11/01/army-receives-first-epir... Don't worry, the US military industrial complex has us covered. We just shoot as much wireless energy as we can at them until their insides melt or something like that. I would be surprised if these things aren't covering the border within the next year or two. Bonus points: uses the same new cool GaN tech that all the new Anker chargers do.

I get the motivation here, but it seems like the people who came up with the "solution" are using an outdated mindset vs those who came up with the "problem".

One, large, central weapon against a swarm of hundreds of small ones? It's a deadly assumption that an entire swarm would come from one direction, for example. Or that it would appear all at once. I'm sure volumes will be written about the strategies yet to come, but if what we've seen in the 20 years of occupation of Afghanistan and Iraq is anything to go by, asymmetric warfare is very effective at tying up resources.

Watching what the twichy youngins do with FPV drones gives me the heebiejeebies -- this has already transferred to battlegrounds in Ukraine, with some serious impacts (no pun intended).

High energy weapons and EW are the only realistic way to deal with it without slinging 100k or 2m dollar missiles around. One problem is these missiles still have debris, though less deadly than an armed drone, they can still be hazardous. Lasers and other high energy weapons are probably the most cost effective and reliable solution for sub 10 pound drones. You generally need only knock out the MEMS gyro, overheat a motor or blind its sensors.

The problem is that drones are cheap and easy to slip around defenses. The other issue is weapon system reliability... the US military expects an extraordinary amount of reliability from munitions / weapon systems and they are supposed to be near 100% reliability. When they launch a missile, it's unacceptable for the missile to not launch.

The Coyote[1] is an interesting counter UAS concept and could be much more cost effective solution.

[1] - https://en.m.wikipedia.org/wiki/Raytheon_Coyote

What about classic flak cannons operated by AI, or things like Phalanx CIWS ( https://en.m.wikipedia.org/wiki/Phalanx_CIWS ) or various automated systems that fire proximity fused ammo?

Just speaking from the US side of things, there aren't any 20/30/40/80/120mm AA rounds being produced in mass quantity and it is more of a scaling issue. Systems like CIWS are also incredibly bulky, requiring a few semi trucks worth of equipment to operate.

There are some German AA guns that have been used for counter UAS, but their export laws are a fucking nightmare to deal with if you ever want to sell, give or trade those systems later on.

I think there's a lot of stuff that's really hard to defend against once it's deployed, launched, or implanted. Think about small arms, artillery rounds, dumb bombs, land mines, or improvised roadside bombs. You have deception, concealment, cover, armor, reactive armor, jamming, and some close-in weapon systems but the best defense is probably more active attacks against earlier links in the kill chain. I think what we (US military, intelligence community, and law enforcement) learned during the "Global War on Terror" (specifically about improvised explosive devices) was to analyze the networks and target the planning, recruiting, training, financing, manufacturing, and distribution components.

War is a game of logistics. Even Sun Tzu's Art of War is all about explaining the logistical realities of war to the aristocrats back home who were making the decisions.

I think there are enough limitations with the current tech that this isn't a big concern at the moment, otherwise we would certainly be seeing these types attacks in Ukraine. But the field is evolving incredibly quickly and it's possible this becomes viable in the future.

Some current drawbacks I can think of:

* Small drones carrying a load generally have a short range, which could make launching such an attack (and surviving the response) difficult.

* Coordination is hard. Manually controlled drone frequencies can conflict. There have been some automated sky displays that went well, but attacking a busy hardened target is much more difficult and I don't think the autonomous software is good enough yet.

* Cost in terms of money, time, and risk. Jerry-rigging drones with improvised explosives is time consuming and dangerous, plus 100 drones cost money.

* Jamming can protect a specific area from drones

I think the coordination problem is pretty much solved. https://www.youtube.com/watch?v=ZbCR8mOPkuo

>Small drones carrying a load generally have a short range, which could make launching such an attack (and surviving the response) difficult.

I understand why such things are relatively rare, but its always sort of a wonder to me that fixed wing or tilt-rotor drones are not commonly suggested for deliveries of this kind. Obviously the flight plan is necessarily more complex when you have to have take-off and landing phases with horizontal components (which requires more pilot training), and the mechanical complexity of a tilt-rotor affair is substantial, but it seems like significantly improving the range could be a significant win. Are the training and design challenges the limiting factor? Is the maneuverability trade-off too much?

Anybody working in this space able to comment?

Honestly, I think it's about the immediate availability, and hackability of the tech. The Ukrainian (army? partisans?) had to respond incredibly quickly to overwhelming force. Off the shelf DJI quadcopters that just happen to have a light control that could be hacked into a release switch + some 3D-printed parts for said switch and fins to convert mortar shells / hand grenades into droppable bombs -- that's one hell of a trial-by-fire hackathon, grab what you can access and convert it into what you need.

Control signal wise, you can go for 100+ km on hobby grade batteries and a simple fixed wing: https://m.youtube.com/watch?v=CYJ2UOrlXgM

The set up would probably cost you around 500 USD: $50 for the frame, $150 for the electronics (gps + fc + 2 motors + 2 servos), $50 for a basic analog fpv system, the rest for the batteries.

https://m.youtube.com/watch?v=RabFZzRyZo8 makes some interesting vtol experiments and his code is mostly open source, on GitHub.

All of these are solved or solvable problems and aren’t impediments in practice.

For comparison: a single unguided artillery shell costs $5,000. A guided shell starts at around $20K and fancy long-range versions can exceed $100K. It takes on average 2-15 shells per kill.

If you had a > $100K budget “per attack” and were making drones in large volume production, then you can easily beat the costs of traditional weapons.

Drones are cheap and can deliver their payloads with incredible precision. Not “150 foot radius accuracy”, but “aim of the left foot of the soldier” accuracy.

It’s a whole new ball game.

Some sort of lightweight material/string that gets caught up in the blades might work. Use it in flak guns and have it float down to the earth saturating the airspace. You could also use it in a barrage balloon manner.

Nets are also a good idea, there has to be a way to physically deny the drones from airspace.

check out what epirus is working on, it's basically a directed HERF weapon that can mount on a vehicle or ship

https://www.epirusinc.com/counter-electronics

Does this work if the drone(s) are automated and don't need an operator or to phone home?

this isn't signal jamming, it's a directed energy weapon. it disrupts and degrades the electronic components that operate the drone.

These weapons are the next generation of targets. Guys driving those trucks are going to get really nervous when HARM drones show up.

cat and mouse is the nature of warfare. They're already working on mounting these types of weapons onto drones as well.

this is a major business line for Anduril https://www.anduril.com/capability/counter-uas/

Aegis go brrrrrrrrrrrrr.

Reply on news.ycombinator.com