Attack of the Drones Revisited
The future of warfare will not have many front line military personnel
Back in 2022 I wrote this post about how drones were in the process of revolutionizing warfare. I recently reviewed it because the Ukrainians have continued to develop drone warfare - as have the Russians though with somewhat less success.
It stands up pretty well in broad strokes but I was wrong in a few details and I missed some developments almost entirely such as naval drones. So let’s start with them.
Naval Drones
Ukraine’s successful use of missiles as well as surface and underwater drones has driven the Russian navy back to the Eastern Black Sea port of Novorossiysk and effectively seized control of the Black Sea from Russia and naval drones have even successfully made attacks on ships in Novorossiysk. The result is that the Russian Navy has gone from controlling the Black Sea to hiding in ports and only moving cautiously between them. This has been achieved despite Ukraine not really having a navy at all as it is all due to missiles and drones.
In some ways, naval drones are not new. In WW2 the Japanese used kamikaze speed boats loaded with explosives to attack Allied naval vessels. In some cases the captain would leave the boat after pointing it in the direction of the warship target, in others he stayed on board to the end. Other things that look a little like naval drones have also been seen. Mines that move a bit or mines that launch torpedoes when they detect a ship near by have similar effects to naval drones. The difference between these weapons of the past and current naval drones is the accuracy and distance/course they can navigate autonomously. As are some of the odder combinations such as a reports of a drone acting as missile launcher to penetrate anti-ship/anti-torpedo defensive nets.
Naval drones are a major threat to all shipping in a way that drones on land are less so because ships are inherently visible and relatively slow. Yes some warships can do speeds on the order of 100km/h though most cannot and no oil tanker or container ship can do anything like that speed, but drones can. Drones are also generally small and hence hard to spot, particularly at night. The Ukrainians have already attacked one of Russsia’s oil tankers and I would assume that at some point Russia’s oil exports via the Black sea are going to stop because Ukraine will successfully sink multiple tankers near one or more of the Russian oil terminals. However that’s just the likely impact drones have on the current war. They actually have a global strategic impact too.
The end of “Fleet in Being” doctrine
Ukraine’s drone naval success will have profound implications globally. Yes it is true that the Russian navy is only medium competent and Turkey’s control of the Dardanelles means Russia cannot reinforce its Black Sea fleet but drones are demonstrably able to do to fleets what aircraft did to individual battleships 80+ years ago. Fundamentally the “fleet in being” doctrine that says that even a fleet that remains mostly in port is a threat which has to be countered by another fleet is gone. A flotilla of naval drones can bottle up a port at the cost of maybe one of the expensive ships in it. Yes theoretically you can send aircraft (e.g. ASW helicopters or even airborne drones) to attack the drones but the obvious response to this is AA missiles on (some of) the drones to take out the attackers. Drones are generally small (compared to warships at least) and constantly moving so hitting one is not that easy and that assumes you can spot them in the first place - tricky if they are small, generally non-metallic and so on.
Other flotillas can force a fleet out of port to retreat and not manage anything offensive. How? Fleets out of port have limited stocks of weapons and a swarm of drones can easily force a fleet to fire irreplaceable missiles at the drones. Eventually (and in some cases eventually is only after a couple of attacks) some ships in the fleet are down to close in gun fire as their only defense and it is easy to see how sufficient drones overwhelm those ships and then continue the process on the rest of the fleet. It must be noted that resupply at sea requires convoys of supply ships that also have to be protected against drone attacks.
A fleet that is stuck in port behind harbor walls, booms and nets for fear of attack as soon as it sorties is not in fact a threat that needs a major counter-threat. And the fact that Ukraine’s exports from Odessa are about the same as they were pre-war shows that drones and missiles are all you need to do that.
One imagines that the Taiwanese (to pick a country completely not at random) are investigating this because a great way to defend against a PRC invasion would be large numbers of drones that can attack the invasion fleet. In fact both parties in that potential conflict should be looking at the prospect (and probably are) because drones will make blockades of ports a lot less risky. The mainland could use drones to interdict Taiwan’s cargo ports and the Taiwanese in turn could do the same to the PRC. You are invited to imagine what happens when a container ship like the Dali or an oil tanker has its engines / rudder taken out by a drone as it turns in to a harbor and smashes sideways into a harbor wall. Since both rely heavily on imports of fuel (coal/oil) the side that figures this out first will be able to cripple the other’s economy for cheap. We know that the Iranians and their various proxies are doing the same too, witness all the Houthi attacks in the Red Sea.
There are obviously counters to naval drones, the most obvious being hunter-killer naval drones that take out the attacking ones far enough away that no harm is done, but this requires investment that no one seems to be doing now. If I were, to pick an example not at random, the Japanese MSDF I’d be diverting a chunk of money to local ship builders to come up with fishing boats or speed boats with modifications to make them autonomous and armed. Note that Japan is one of the few countries in the world that has market leading companies in shipbuilding of everything from 12 ft day fishing boats to monster 400m container ships, other countries (e.g. USA, Britain) no longer have this knowledge or have never had it in the first place.
Airborne Drones
When I wrote the original piece was thinking more of flying drones and in general smaller versions of them such as the short range quad-copter style ones or slightly larger winged drones like the Russian/Iranian Shahad/Geran ones (~ 2.5m x 1.5 - 8ft x 11 ft) but those are not the only types of drone Ukraine has developed. For example, in the last couple of days Ukraine deployed a large drone (the cessna-like A-22 Foxbat aircraft converted to remote control) to hit Russia's drone factory in Tartarstan and in the past this and other larger drones have attacked numerous Russian oil refineries.
It also launched swarms of drones of up to 50 (each) to attack 4 airfields in Russia destroying (or damaging enough to make unflyable) at least a dozen Russian aircraft. These are probably mostly smaller drones, although the attack on Engels base (over 750 km from Ukraine) probably used fewer, larger ones. Likely cost of the 50 drone swarm around US$1M. Cost to Russia between $100M and $1B (and in fact priceless because Russia literally cannot replace these aircraft in quantity).
HI Sutton has a good page describing all sorts of Ukrainian drones that is worth referring to for more details.
I should note that here’s a place where I got the details slightly wrong. When I wrote the original post almost 2 years ago I expected most drones to be reused and relatively few to be OWA (one way attack) kamikaze drones. It seems that jamming by both sides means that drones of all sizes and missions are now mostly one time use. But this is not so important because the cost of each drone is dropping fast (that bit I did get right)
For example there's this plastic pipe done (Forbes article - archive , also discussed on this X thread). The drone costs under $10,000 in materials (probably half that) and is made in part of standard plumbing/electrical pipe and plastic water bottles as fuel tanks. It can be constructed in a small tool shop by a couple of people in an hour or so assuming the key bits like the motors are available so. It has an estimated range of 200-300km and delivers perhaps 3kg (7lbs) of explosive payload in addition to the molotov cocktail from unused fuel. I suspect that the mostly plastic body means that this drone has a relatively small radar cross section and hence is almost invisible on night time attacks
A single drone of this sort probably won’t do a huge amount of damage, but as part of a swarm it is rather more of a threat, and if it can be aimed well it could certainly do major harm to critical infrastructure like electrical transformers or oil/gas pumps. A single blown up transformer or pump is easy to replace, hundreds of them are not.
The key to this drone and to other related ones like this slightly larger jet-powered Shahad equivalent, is that they need expensive and complex defenses to be stopped and that once they have got past the front lines such defenses are few and far between. Sure Russia can protect airports a bit, and it can protect oil refineries a bit but those defenses are both limited and struggle to protect the whole area against multiple simultaneous attacks. Ukraine has already sent drones to Moscow, to St Petersburg and to all sorts of places nearer to the front lines such as Rostov. If Ukraine can deploy enough of these at a time and enough in total they can disrupt the Russian war effort and perhaps in the process disrupt Russian civil society enough that it collapses. Cruise missiles like the Storm Shadow or Russian Kalibr inflict lots of damage when they hit, but they cost millions of dollars each, are not always accurate and get shot down sometimes anyway. You can get hundreds of pipe drones for the same price as a single missile and as the 50 drone swarms on Russian airfields show, they can be just as effective.
Drones on the Front Line
One of the differences between drone warfare and what happened before is that drone attacks tend to be slower. In prior wars if you took a shot at something and scooted away you’d likely survive the return fire (or if you didn’t your friends found out a few seconds later that you hadn’t ducked properly). With drone warfare that’s not the case. Drones can hunt you down and chase you as you run away.
https://twitter.com/casualtanker23/status/1772851022601576728
Now you might think that this slower response means drones can be easily shot down, but that seems to be only partly true. Both Russian and Ukrainian ground forces are strongly motivated to find ways to shoot down drones and clearly manage to shoot down a lot of them, but a lot of them is not all of them and the ones that get through are highly destructive. So much so that both sides now try and hunt other drone operators (e.g. this forbes article, archive) because that’s more effective than shooting down the drones themselves.
The key to drone success appears to be the ability to fly lower than many defensive systems can handle, to jink unpredictably, and to be relatively small so targetting has to be very accurate in the short period of time that the drone is in view. Yes the US and other advanced nations have weapons that can shoot down drones. Those weapons tend to be very expensive and limited in quantity. So yes they shoot down the first few drones and then they run out of reloads and the next drones get through. We may already be seeing this with the Houthis in the Red Sea, we have certainly seen in on land where Saudi Arabia has used patriot missiles to defend its oil infrastructure from drone attack with only moderate success.
I expect to see a lot of work on C-RAM systems in the next few years but C-RAM and friends only work where they are deployed. On the front lines C-RAM is a good thing, and C-RAM systems must be dceployed there to take out as many drones as possible but that’s not a complete solution. Drone warfare seems to work best by breaking through and then finding vulnerable bits of the supply chain to attack. You can’t have a C-RAM to protect every industrial estate, every transformer, every pumping station. You probably can’t even have one for every ammo dump, every artillery firing point etc. and if you did one suspects that there would need to be a lot of deconfliction software developed to make sure that C-RAMs don’t shoot down your own drones and don’t mistake each other for threats and/or don’t get taken out by shooting at drones flying between two batteries and so on. If someone can develop a powerful enough, portable enough laser this will be ideal but as far as I can tell we are quite some way from that goal.
Right now though we’re at the cobbling together of bits in the hope that it works stage
The one thing that is clear is that drone dominated battlefields are generally bad places for fighters who cannot easily defend themselves against drones. The Russians have, at various times, tried mass wave attacks which do potentially exhaust the capacity drone operators to respond, but those attacks do tend to encounter artillery. And in the reverse the Russians have mined many sections which also work well to stop large assaults. Drones are very good at spotting and picking off individual / small team incursions and/or defenders so as drones become more widespread we are likely to see relatively large areas of no-man’s land between the front lines that are completely bare of humans except when a large assault is happening. Outside of those times the no-man’s land is left to drones and C-RAMs trying to kill other drones and drones trying to evade the defenders and attack.
Another recent forbes article (archive) discusses recent Ukrainian successes against Russian air defenses. The fact that air defenses, which you might think would be able to detect and destroy drone attackers, have problems doing so shows that a lot of work will need to be done to protect potential targets in an era of drones. Drones certainly aren’t invincible but we aren’t anywhere near a comprehensive reliable solution to the threat they pose.
That threat is quite a democratic one (democratric as in the phrase “ware is a democracy, the enemy also has a vote”)
Drone Supply Chains
One thing that the Ukraine/Russia war has made clear is that drones can be made of almost anything. Once you have the basic design smaller drones can be churned out by home 3D printers / light engineering CNC shops and the like with very little advanced hardware. The only two really complex bits are the engine and the controller, both of which these days are effectively “off the shelf” components.
Ukraine may, for example, use Chinesium engines in its current drainpipe drones but that’s merely because they are available in quantity for cheap. Ukraine has used small jet engines and the like from both the UK and Germany and nothing is stopping it from buying motors from other places too if it can’t get Chinese ones. The slightly more secret sauce is the controller but again that doesn’t seem to be a limiting issue. Sure there are dedicated controllers from DLJ and other Chinesium makers that are readily available but nothing in those controllers looks to be particularly complex either and I suspect that the right radio components connected to a Raspberry Pi (or similar single board computer) would be entirely sufficient - presuming you can’t simply reverse engineer the controller board and then build it yourself.
On that note, there have periodically been X-storms of manufactured outrage that Iranian drones have chips from companies like TI in them, but these are pointless. Drones are (relatively) low technology and can be built from components sourced using the standard supply chains used to build any number of personal electronics, toys and tools. That means they will incorporate chips from all over because no chip supplier can track, let alone stop, the path of chips into sub-components into components into final assembled product with sub-components and components being traded via middlemen of all sorts, many of whom are equally unable to work out the end use of the products they are selling. It’s probably only the last stage that arms-trafficking laws would apply and you can be sure that nation states will set up the best shell companies they can to disguise the final destination.
Fundamentally anyone with a few thousand dollars and access to the internet can order the parts to build a fairly good drone. The electric quadcopter FPV drones are in the US$500 range. Those are payload and range limited (3 miles/5km) but
Move it to hundreds of thousands of dollars and you’re looking at multiple far more capable drones. To be honest the trickiest part of a drone these days is the explosive - presuming you want something more than a molotov cocktail - and that’s only an issue if you aren’t a national military / military supplier.
Mind you when it comes to drone suppliers, I find it fascinating that the US seems to be severely lagging with drones that don’t work well in combat.
[M]ost American start-ups' small UAVs have failed to prove themselves in combat, dispelling the companies' hopes that a badge of combat testing would generate sales and attention. It's also bad news for the Pentagon, which needs a reliable supply of thousands of small uncrewed aerial vehicles.
The WSJ noted that US companies still lack a significant presence in the first war where small drones are playing a critical role. US-made UAVs tend to be expensive, faulty and complicated to repair, say drone company executives, the Ukrainians on the front line, Ukrainian government officials and former US military officials.
Drone Evolution
In the last two years we’ve already seen significant drone evolution. Currently it looks like everyone is looking to use AI image processing to help with drone navigation under the assumption that GPS is blocked/unreliable as is communication back to the launcher. A drone that can track main roads and perhaps spot stars and use the rules of celestial navigation the way a bush pilot does is going to be unjammable. Thanks to google earth and other GIS projects it will be possible to simulate an entire journey, pick suitable waypoints and triangulation points for the final attack so the drone can navigate without any need to rely on GPS or other external data.
We’ve already seen the Ukrainian Baba Yaga drone used as a relay point/carrier for smaller FPV drones. I suspect we’re going to see more of this kind of drone combination. A relatively large jet drone (or worse, group of them) could penetrate enemy airspace and get close to the target and then launch a swarm of smaller dones. If those drones have the AI vision concept mentioned above they can autonomously swarm the target and probably observe and then destroy whatever weapon the defenders use to take out the jet drone once it has dropped off its payload of mini drones as well. We’re probably some way away from the SFnal transformer concept where a single object transformes into a swarm but I can see how you could get close to this with a thin skin/exo-skeleton containing many drones within the skin.
Two things that I haven’t (yet) seen (reported) are drones that can loiter on the ground and point-point radio or laser communication with drones.
To take the latter first. With the right accuracy this solves the jamming issue completely and it also strongly assists in the navigation issue because having a beacon on a known heading from the drone makes location calculation easy. Especially if you include timing pulses to measure how far away it is by counting how long it takes to return a signal. There are, of course, issues here in that a laser requires line of sight and radio requires almost the same but absolutely nothing says that they relay point needs to be in the same place as the drone launch point - a static blimp several miles back and at relatively high altitude would be fine. Such a blimp could also serve as an observation point to spot incoming drones too and potentially as a launch point for counter drone missiles/drones when attackers are spotted.
The loiter on the ground concept has a couple of applications. One is as mobile minefield:- fly drones to near a road/railway where you expect some high value traffic, have them land and wait until the traffic is detected and then launch to attack - probably not from the direction anyone is looking at. A second is as a way to take a slow drone a long way inside enemy territory by having it only fly at night and “rest” during the day somewhere. It doesn’t neccessarily need to rest on the ground. On the tops of buildings also works or even on the top of electric pylons or water towers. Sure the drone is visible there if you look at the pylon, but a pylon in the middle of a forest is unlikely to get people looking at it and forests are something that Russia (to pick a country totally at random) has many of. Heck if you are clever enough and the drone small enough it could ride on top of a train and let the enemy do the work transporting it to its destination
A single large slow drone carrying smaller dones that it can drop off to launch later would be able to cause havoc all over the country. A slow drone that made into Russia and then after a pause u-turned and headed back to the frontlines would likely fail to trigger much if any response from Russian air defenses because it would be percieved as coming from a safe direction. At least not until it got close to its target.
I suspect we’ll see these kinds of developments in the near future. None of them require too much in terms of new hardware or software, just a bit of cunning in applying existing kit. I hope we see more innovation from developed nations because right now, apart from Ukraine, the majority of drone innovation seems to be coming from the Iranian/Russian/Chinese axis of evil
But on the bright side, the US is the only country to have male & female & neuter drones, and in different colors as well. That has to count for something, no?
We used to have a lot of small machine shops capable of making items like drones, without a lot of government "oversight". Many of these are gone - maybe thanks to the oversight. Could be with our new socialist govt, we'll be able to use those old-school rocks & slings to fight future wars.
"A fleet that is stuck in port behind harbor walls, booms and nets for fear of attack as soon as it sorties is not in fact a threat that needs a major counter-threat."
Taranto, November 1940, Royal Navy carrier biplane torpedo bombers, Fairey Swordfishes, the same type planes that would disable the Bismarck in mid-Atlantic in May 1941, eviscerate the excellent Italian fleet in port.
Mitsuo Fuchida, "The World At War" episode "On Our Way", notes "we [Japanese] learned very much from the British ability to launch aerial torpedoes in shallow water and used that knowledge to plan our Pearl Harbor attack".
The offense/defense cycle has been going on for some time now.