View in browser
Tech Surge of the SMO: AI, Drones, EW, Countermeasures, and More of the Latest Advancements
Simplicius
Jan 4
∙
Paid
READ IN APP
The following is a premium article of a hefty ~4800 words in size, covering the latest technological developments on the frontline in the area of drones and AI tech in particular. The report is packed chock-full of exclusive videos and hard-to-find details that you won’t see anywhere else, which I’ve collated through tirelessly poring over obscure sources and channels. So if you’re interested in the technological aspect of the Ukrainian war in particular, this is another segment you don’t want to miss.
We haven’t had an update on the state of the war’s technological progression in a while and the new year brings the perfect time to do so. One of the reasons for that is because there have been a lot of predictions on the sweeping changes said to take root on the frontline by 2025, and so it’s appropriate to discuss how close these projections have been.
One of the notoriously contrarian predictions from many months ago was that 2025 would actually see not the supremacy of drones and FPVs, but rather their negation and decline. A French army chief made this curious claim in June of last year:
https://www.defensenews.com/global/europe/2024/06/19/small-drones-will-soon-lose-combat-advantage-french-army-chief-says/From the article:
The advantage now enjoyed by small aerial drones on battlefields including in Ukraine is but “a moment in history,” French Army Chief of Staff Gen. Pierre Schill said at the Eurosatory defense show in Paris.
While anti-drone systems are lagging and “leave the sky open to things that are cobbled together but which are extremely fragile,” countermeasures are being developed, Schill told reporters during a tour of the French Army stand at the show June 19. Already today, 75% of drones on the battlefield in Ukraine are lost to electronic warfare, the general said.
One must wonder if the French army chief knows what he’s talking about; the article further points to upcoming French vehicles which may include anti-drone ‘missiles’ and ‘40mm airburst grenades’. But these will prove useless against FPVs which are much too quick, ubiquitous, and undetectable in the frenzy of combat to really be reliably destroyed by such expensive countermeasures.
To prove his point, the French army chief compares FPVs to the Bayraktar drone, saying they will disappear just as unceremoniously from the field:
First-person view drones currently carry out about 80% of the destruction on the front line in Ukraine, when eight months ago those systems weren’t present, according to Schill. The general said that situation won’t exist 10 years from now, and the question could be asked whether that might already end in one or two years. Schill cited the example of the Bayraktar drone, “the king of the war” at the start of the conflict in Ukraine but no longer being used because it’s too easy to scramble.
But the article does make one powerful point: that Western armies are essentially paralyzed from committing too directionally in one weapons program because the possibility is too high that the albatross program can be obsoleted by a new development very quickly:
The pace of military drone development means that Army can’t commit to large buying programs, because an acquired capability can become obsolete in five months, according to the general. Schill said today’s drones fly better than those two or three years ago, with more computing power onboard that is capable of terrain-based navigation or switching frequencies to escape jamming.
Everyone is now focusing on one magic silver-bullet system to take on drones. But the real answer lies in a total, holistic approach with the understanding that losses from drones will simply become an inescapable reality of modern war. This is how Russia has now chosen to approach the situation, simply mitigating drone advancements as much as possible not with any one particular system aimed directly at combating them, but rather through the total synergistic strategic realignment of the armed forces as a whole. This includes everything from surveillance, EW systems, the tightening of the entire operational decision tree and OODA loop, direct personnel training, anti-drone prophylactic systems for vehicles, but also the actual combat tactics and strategies employed, like Russia’s now-famous ‘dispersion’ approach, better known as the ‘death by a thousand cuts’. In an interview last month, a Ukrainian soldier had remarked how it has become very difficult to hit Russian troops with drones on his front due to the ‘slow trickle’ method they’ve begun to utilize in accumulating at a forward position. When there are only tiny groups of two or three men at a time weaving into the position from a variety of random directions, the AFU drone teams become dispersed and paralyzed from lack of concentrated targets.
Ukraine’s famed drone king “Magyar” had quite the opposite prediction in September of last year, stating that by March 2025 drone pilots would already be old news:
https://www.unian.ua/weapons/bezpilotni-droni-madyar-zayaviv-shcho-piloti-droniv-vidhodyat-u-minule-12760995.html"Currently, hundreds of artificial intelligence systems are being developed simultaneously, and they are being tested in experimental modes. After six months, the pilots will no longer be required. You will need people who will simply lift the drone a meter above the ground. And then the drone itself, depending on its development, will decide what to attack, how to distinguish a Zhiguli from a tank, and will definitely not confuse the Ukrainian with the enemy, "Magyar said.
This seems a bit premature, after all we’re already nearing his six-month limit and the battlefield is not overrun with artificial intelligence systems. But there has been more and more noise in this direction.
For instance, top Ukrainian electronic warfare expert Serhiy Flash found an uptick in Russian forces using AI targeting drones in the Kursk region, even showcasing pictures of the recovered electronic boards:
Russian drone with target acquisition and auto-following from the Kursk bridgehead.
Recently, there have been more and more such drones. Holding a moving target is far from ideal, but it works.
I remind you that a drone with auto-target acquisition completely neutralizes trench electronic warfare.
In the case of mass production, the auto-capture module increases the price of the drone by 100-200 dollars
Meanwhile, Russian forces on the other hand recovered some of the Ukrainians’ own parallel efforts, likewise showcasing a special Google AI CPU.
Report:
Ukraine and Google artificial intelligence!
Recently, the wreckage of a Ukrainian quadcopter (FPV) found on the battlefield uses an artificial intelligence (AI) control system. After opening the quadcopter, it was found that the Ukrainian uses the Edge TPU development board developed by Google.
The Edge TPU board is the computing core unit of Google's Coral platform, which can be purchased publicly at a price of about $130, and Coral is a platform that provides complete hardware and software solutions for artificial intelligence. Unlike GPU boards, TPU boards are much more optimized for large-scale parallel computing required by networks.
Eric Schmidt, the former CEO of Google, previously said that because of the war in Ukraine, he is now a licensed arms dealer! which aims to help Ukraine achieve artificial intelligence technology. He also believes that the US military should eliminate useless tanks and replace them with drones equipped with artificial intelligence.
Ukraine hopes to use artificial intelligence-equipped drones on the front lines to help the country overcome Russian jamming systems that have become effective and enable drones to operate in larger groups. Since both sides of the Ukrainian war are using electronic warfare systems that can disrupt communication between the operator and the drone, the hit rate of FPVs has decreased.
At present, most FPVs have a hit rate of 30-50%, and the hit rate of novice operators may even reach 10%, but it is said that the hit rate of AI-controlled FPVs in the future could be as high as 80%.
Here are two such tests from Ukrainian units of auto-tracking FPVs:
And one actual hit on a Russian tank which appears to be using an automated AI tracker:
Here is the same Ukrainian drone master ‘Magyar’ who goes in detail about a captured Russian drone with AI “machine vision”:
Germany is slated to begin supplying thousands of AI-equipped drones to Ukraine as well:
Germany will massively supply Ukraine with kamikaze drones, which are called "mini-Taurus", - Bild .
▪️We are talking about Helsing company's shock BpLa, equipped with AI technologies , which give them higher autonomy in the conditions of EW means of operation;
▪️The flight range is allegedly 4 times (!) greater than that of typical Ukrainian kamikaze drones;
▪️They are compared with Taurus long-range missiles, which Scholz refuses to provide;
▪️U companies they allegedly ordered 4,000 such drones for Ukraine. Batches of several hundred per month may begin to arrive from December.
Besides that, there has been a series of reports regarding other major ‘shadowy’ American defense and AI labs using Ukraine as a test bed with increasing scale.
The Pentagon’s infamous project Replicator has reportedly been at the forefront of this:
https://www.defenseone.com/technology/2024/11/newest-replicator-drones-proven-battlefields-ukraine/400997/Replicator acquired and tested several Anduril systems. One of the most interesting and promising directions has been in AI swarm tech to combat EW jamming—from the article above:
For instance, the source said, Anduril has developed flying mesh networks to enable drone swarms to swap data even amid heavy electromagnetic warfare interference, “relaying that data along multiple UAS[es] so that they can have long data-chain links.”
The drones' high level of autonomy also helps them evade EW effects and interceptor missiles, the source said.
“Let's say that I'm relaying comms, and then all of a sudden, the Russians pop up an EW bubble. The drone can say, ‘OK, I expected that. I'm not even going to go to my fallback positions. I'm gonna fly over here. I'm gonna go into a place where their jammer—which I can triangulate—is not affecting my link anymore.’ So you get all these networks that physically reconfigure their geometry to be robust to jamming.”
Imagine a swarm of drones all interfacing their data together: when one drone falls into the cone of a disabling EW beam, the other drones can triangulate the affected drone’s position and send it appropriate waypoint data to get to reposition itself to safety.
In fact, OpenAI has now announced not only a partnership with the Pentagon in developing warfare AI but has also appointed Paul Nakasone to a top position within the company; Nakasone is former head of the NSA and US Cyber Command.
Here’s the President of OpenAI, to show you the kind of dark directions the company is headed into:
One of the chief issues the US and partners have gleaned from the Ukrainian war is the sheer importance of mass scale at a low cost. For example, the drones being produced for the US Marines turned out to be an outrageously unsustainable $94,000 each—this is just for the basic DJI Mavic reconnaissance analogues:
So now the Pentagon and DARPA have announced new programs to address this, which also include the concept of highly versatile and customizable production infrastructure which can redistribute manufacturing in order to lower the price of mass ballistic missile strikes on a large ‘centralized’ production hub:
The Pentagon's Advanced Research Projects Agency DARPA announced the concept of rapid deployment of military production within 2-3 days at the Freedom Forge 2.0 seminar. The concept is clearly borrowed from Ukraine and the" Army of Drones " of the Armed Forces of Ukraine. These should not be large factories that take months or years to build, but a network of decentralized, small, flexible production facilities with the ability to deploy anywhere.
Instead of machines, robots and 3D printing will be used. All the necessary equipment is planned to be manufactured on-site. Moreover, everything will be produced - from a drone to the barrel of a tank gun.
The DARPA initiative points out that the United States does not ignore the "industrial" component of the experience of the SVO and seeks to play ahead of the curve, coming out to the next conflict more prepared than any likely opponent.
In short, they want “pop up” factories that can be quickly relocated and set up to run as autonomously as possible, with highly modular capabilities of producing all kinds of different weapons systems. This is extremely ambitious to say the least.
The US also sees the success of the ‘mosquito fleet’ model against standard legacy surface fleets which are ‘giants with feet of clay’ against the modern drone-centric evolution:
NATO plans to create a fleet of sea drones to protect infrastructure in the Baltic and Mediterranean Seas
▪️This was announced by NATO Supreme Allied Commander Transformation Pierre Vandieu in an interview with Defense News.
▪️Drones would allow NATO to “see and monitor the situation on a daily basis.”
▪️They want to complete the project by June 2025.
RVvoenkor
Every nation continues to conduct various drone swarm tests
https://breakingdefense.com/2024/07/swarm-wars-pentagon-holds-toughest-drone-defense-demo-to-date/China increasingly holds drone exercises to incorporate assault and recon drone usage down to the squad level:
But one gets the impression that Russia and Ukraine are both leagues ahead of the countries on the sidelines for the simple fact that real battlefield conditions create exceptionally nuanced use-cases and requirements for ad hoc improvisation that are simply impossible to anticipate or test for in ‘exercise’ conditions or as part of procurement proposals for these various labs like Anduril.
I’ll give a few examples of this.
For instance, both the Russian and Ukrainian sides endlessly modify and re-modify virtually every component on both the receiver and transmitter side of the drone equation, creating complex forests and hierarchies of electronic wizardry to deal with constant unexpected daily challenges, evolutions, and advancements.
Recall long ago both Russian and Ukrainian sides began to utilize a system of drone-dropped “nav” signalers on the ground—a kind of electronic waypoint beacon—which can direct drones to various waypoints in heavy EW-contested environments. Now, Ukrainian units report the advent of a new system on the Russian side:
These are FPV “traffic lights” being put up around various Russian main supply routes and arteries which show three lights like red, yellow, green that warn:
No FPV
FPV far away
FPV nearby
Thus, as you’re traveling along the road, the light with its automatic detector can immediately warn you when you’re potentially being hunted.
One Ukrainian expert’s reaction:
I was even surprised that my information about the enemy's anti-drone traffic lights caused such a reaction in the media.
What information is there and my thoughts based on it.
Putting a FPV signal receiver in each traffic light is expensive and risky, such an autonomous traffic light will be quickly stolen by the Russian military.
Therefore, the implementation is as follows: somewhere in the center there is a FPV signal receiver at a high point and under guard.
The traffic lights are spread out around the district and have a LoRa connection with the central control point. Thus, the traffic light itself is a cheap product that does not represent value.
So far, I do not know whether the information comes automatically from the receiver, or is processed each time by a person who, based on the FPV picture, makes a "friend or foe" conclusion.
Do we need this on front-line roads? What do you think?
Russian units also began to install video interceptors in their vehicles that intercept the unencrypted FPV video signals from enemy drones. Here is one example of a Russian official whose car has a built in screen with such an interceptor. Upon driving in a ‘hot’ area, the video receiver picks up a Ukrainian FPV hunting the road somewhere just above him, causing him to immediately back away and scramble.
As you can see, though, someone didn’t get away.
Another innovation on the Russian side is a networked series of solar-powered microphones which triangulate Ukrainian drone incursions, distributing the data to nearby brigade HQs:
Along the entire border with the Russian Federation and in the rear, our enemy uses a network of observation points for Ukrainian UAVs
based on special "microphones".
Such points are mounted on poles and mobile communication towers. They include 4 microphones, which, when recording sounds from UAVs, transmit information via radio channel or via mobile communication network to the command post.
Such an observation point has an autonomous power source and a solar battery.
Now with the advent of fiber-optic cable drones which are immune to EW jamming, new innovations have to be constantly thought up. Such a one on the Russian side are new strobe jammers which target the drone’s video camera sensors.
An example below, but warning for epileptics—major strobing lights:
Ukrainian units write of many other innovations on the Russian side. For instance, recently even Russian medium-class recon drones like Orlan, Zala, etc., have begun equipping video signal jammers of sorts. When Ukrainian FPVs get close, the Russian drone jammer automatically detects the frequency of the FPV’s video channel and then pumps that same frequency back to it, but at a much stronger dB. This overwhelms the FPVs video channel, causing it to go blind.
The impressive innovation here is that this detection, signal analysis, and redirection is all instantly automated on a small jammer placed on the winged recon drone.
That’s not to mention that Ukrainian units have been finding all kinds of new “interesting” things attached to Russian recon drones, like rare L-band radars to detect Ukrainian radar emissions along the frontline:
In the photo, a rare Orlan. It does not have a camera, but with its receiving antenna it searches for signals in the L range of our counter-battery radars and mini radars of a not very high frequency range.
And the last type of Orlan I have never come across. This is a flying VHF direction finder. I read that it exists and there are terrible rumors about it in the ZSU (they will find us all from the sky), but I have never seen it as a trophy, which means there are few of them, which means it is not so scary.
Both sides continue pioneering the “mothership drone” class, with Russian claims that new ones can extend FPV range to nearly 80km. That’s because the mothership first ‘ferries’ the FPV to long distance while acting as its signal repeater, then drops the drone over valuable targets, after which the FPV is remotely controlled by an operator.
Here’s a Ukrainian one in action:
Meanwhile Russian helicopter crews now use FPVs to take out Ukrainian naval drones:
If that wasn’t bad enough, those very same Ukrainian “Magura” naval drones have just achieved a ‘historic’ first of reportedly taking out two Russian Mi-8 helicopters over the Black Sea via guided Soviet R-73 air to air missiles adapted to be fired from the drone:
FighterBomber appeared to confirm the shoot downs.
That’s not to even mention those very same Ukrainian naval drones have already been armed with machine guns and were seen firing on Russian helicopters several weeks prior to this.
Ukraine also claimed to have conducted its first ever UGV or ground robot assault on Russian positions:
Meanwhile, Russia has begun mass producing the first fiber-optically controlled UGVs:
Both sides are beginning to mass produce all kinds of UGVs, here’s a recent Ukrainian batch:
Russians are even beginning to roll out EW-carrying UGVs to protect troops:
On the contact line, the most important task is to provide electronic cover for assault groups, ensure the safe evacuation of the wounded from the front line, as well as military equipment and ground-based robotic complexes. Among the new products developed for this purpose is the tracked jamming robot for covering assault groups "Reb Wall-e", which in August was first presented at the International Military-Technical Forum"Army-2024".
On this mobile platform, electronic warfare equipment is installed — jamming stations "Fumigator-FPV" and "Fumigator-UAV". They provide a circular cover from unmanned vehicles, including FPV drones: the first one suppresses enemy UAVs, the second creates a protective dome with a radius of about 150 meters.
And now even the heavy class ‘Baba Yaga’-style drones are coming in un-jammable fiber-optic varieties—note the big black cylinder underneath, where the huge fiber spool sits:
Similarly, Russian Kalashnikov Concern is testing a new tethered drone for prolonged sector surveillance as a kind of replacement for aerostats:
Russian brigades are designing their own DIY wearable helmet-mounted jammers:
The 114th brigade has its own helmet-mounted electronic warfare system! Report by Sladkov.
The opening of the video shows a Ukrainian FPV falling limp to the ground behind a soldier armed with the homemade EW helmet:
Before you express shock at the likely brain damage incurred from the signal, note that they state it should only be turned on when a drone has already been detected rather than running hot at all times.
In light of all the above, the transformation slowly taking place within the Russian armed forces is a holistic one under Defense Minister Belousov’s initiatives. The aim is to integrate all the various systems in an organic way through the different levels of the military structure such that the smoothly networked operation of these assets can effectively be carried out between various involved units.
An example is this new photo series from a Russian artillery academy which shows the cadets training on integrating various drone sensor data with networked battlefield ‘awareness’ software packages:
Footage of the educational process at the Mikhailovskaya Military Artillery Academy from the kindergarten "Military education. Taking into account combat experience."
Cadets are trained to work with artillery using quadcopters and various software, including the Glaz/Groza software package , designed to transmit information from UAV operators to tank crews, artillery crews and command posts.
To train the crews of the Msta-S self-propelled guns, the academy uses software for the Dilemma simulator using virtual reality technologies .
In the training of future leaders of the operational-tactical level, various software and experience of the SVO are also used, including that personally acquired by the officers-students of the academy.
"Currently, a non-member of the SVO cannot enroll in a master's program. Therefore, they themselves make their proposals on what to change in the educational process, what to give them, what they know and understand. And the program for them, naturally, also changes," - Deputy Head of the MVAA for Academic and Scientific Work.
Military Informant
Systems akin to this, demonstrated previously:
So, what does the future hold? More autonomous systems trickling in as prices for customized AI chips drop, as explained in this WSJ article from a month or two ago:
https://archive.ph/KKMTBIt states:
Now, an even bigger breakthrough looms: mass-produced automated drones. In a significant step not previously reported, Ukraine’s drone suppliers are ramping up output of robot attack drones to an industrial scale, not just prototypes.
Enabling the upshift is producers’ successful integration of inexpensive computers into sophisticated, compact systems that replicate capabilities previously found only in far pricier equipment.
“None of this is new,” said Auterion founder and chief executive Lorenz Meier. “The difference is the price.”
Kyiv is set to receive tens of thousands of Auterion’s miniature computers, known as Skynode, which should hit the battlefield early next year. Vyriy Drone, a top Ukrainian drone startup, said it would produce several thousand autopilot drones starting this month. Other companies are also ramping up production.
The problem with this is the current phase is still ‘terminal guidance’ only, rather than AI-driven “free range” hunt mode. That means a human operator still has to find a target first while the AI only takes over the last few hundred meters, should the signal get jammed.
Using terminal guidance overcomes those issues. Autopilot mode can be engaged roughly two-thirds of a mile from a target—well outside the short range of jammers. Drones with autopilot can strike objects behind hills as they don’t need to maintain a signal with the pilot in the attack phase.
As such, the latest AI ‘targeting’ FPVs are not as groundbreaking or ‘game changing’ as some would like because Russia has now adapted in increasing its force distribution to such an extent that doing the initial target acquisition is already the hard part. Russian troops now regularly trickle into positions a few at a time and by completely irregular means:
This puts heavy stress loads on drone operators that even AI cannot solve as of yet.
Furthermore, should AI drones begin proliferating there are arguably new jamming methods that could still prove quite effective against them: for instance, the earlier ‘strobe’ jammer. Since the AI targeting mechanism heavily relies on a clean video signal, once an activated ‘dazzler’ begins ruining the cheap CMOS sensors in the drone cameras, the AI algorithm will sputter and be unable to track the target. I can foresee a bunch of cheap strobe lamps attached to soldiers’ backs, helmets, vehicles, etc., which activate upon detection of an FPV nearby. There’s no way for the AI algorithm to bypass this, and the only ‘solution’ would be equipping the drones would far more expensive camera sensors which are resistant to the strobes, which would defeat the whole purpose of mass produced FPVs.
On the other hand, equipping oneself with huge blinking lights obviously gives away your position to every other surveillance drone for miles around—so there are quite some trade-offs.
Ultimately, in my opinion AI has been far more transformative in the ‘behind-the-scenes’ action of the war, rather than the dinky terminal guidance aspect for FPVs. AI data sorting and analytics is the real game changer which has already been used through various Google and Darpa programs like Project Maven to analyze and collate vast reams of satellite data to identify hidden Russian targets. This is where the true revolution is taking place, while the ‘autonomous drone systems’ mostly lag behind.
The reason probably has to do with the much faster advancement of language models which can be used for various data analytics purposes in a far more natural way compared to robotics and such, which is always a generation or two behind.
But once again, the nation which will have the edge in the future will be the one that can apply all of the various advancements in the broadest, most integrative way possible, up and down every echelon of the structure—from the general staff to squad level.
Just two weeks ago Russia had announced the creation of an Unmanned Systems Forces military branch:
https://www.fdd.org/analysis/2024/12/20/russia-announces-plans-to-form-unmanned-systems-forces/The creation of the 'Unmanned Systems Forces,' announced by Russian Defense Minister Andrei Belousov on December 16, follows the formation of several units armed with drones.
Those units require support of the ministry, including training, supply and staffing, Dmitry Kornev, founder of the Military Russia analytical portal, told Sputnik.
"One of the effective ways to solve this problem is to create a separate branch of the military, which will ensure the solution of the tasks that the armed forces face in providing and using drones and unmanned aerial vehicles anywhere," Kornev explained, suggesting that the branch will be part of the Russian ground forces.
As such, Russia is preparing to tackle this very challenge of integrating all the new advancements across the armed forces. Make no mistake, the West certainly has many advantages, and even the lead in certain directions of drone and AI tech, but there are likewise many areas where Russia has a huge headstart, particularly where it is simply impossible to make inroads without first hand battlefield experience in a modern peer-to-peer conflict.
This is why the West has flooded Ukraine with its secret operatives, but even they can only transmit so much latent knowledge directly to hundreds of thousands of their compatriots back home who will never experience it first hand.
But stay tuned for further future updates where we’ll continue to explore more of the latest technological developments from different aspects.
