Lessons from Russia’s Shahed fleets for US long-range drones and air defenses

Poland, with its lengthy history of fighting back against its eastern neighbor's invasions, responded by shooting those drones down, mostly using Dutch Air Force jets in-country under NATO auspices. While this was the right thing to do given the choices available, the question reverberated around NATO: Is this the best we can do?

The question remains extremely apropos and increasingly global. Groups ranging from South and Central American drug cartels to Houthis in Yemen have used drones to compensate for limited traditional air power. Outside of Ukraine, cheap drone kit allowed Islamist rebels to charge into Damascus and drive out Bashar al-Assad, whose bloody reign seemed to have hardened into a lifelong scab on the Middle East almost a decade ago.

But despite a number of clever Ukrainian drone applications, Iranian Shaheds and their Russian copies — Geran and Gerbera — are undoubtedly the best unmanned system to have emerged since the start of the war in Ukraine. Indeed the U.S. and Ukraine are both busily developing their own imitations.

Over the past nearly four years it’s already become a cliché to speak of Ukraine demonstrating that modern warfare can break beyond the sorts of precision munitions that the U.S. defense industry has spent the last 50 years fine-tuning and marketing to their foreign allies. Mass once again matters. It necessarily follows that the need to curtail the oligopoly of U.S. defense contractors is urgent, given their practice of demanding decades of government funding in order to develop exactly the types of product that they can sell to the government for the highest mark-ups.

While both [phenomena above — the return of mass matériel and the reluctance of American weapons makers to produce it in earnest — are true, they are hardly the only lesson that Shaheds present for the West’s would-be producers of deep-strike drones or their countermeasures.

Modularity and the trouble of Trojans

Ukrainian intelligence routinely drops announcements about some brand-new Shahed drone that’s threatening to take over the skies, whether it’s outfitted with new AI chips or jet engines.

Russia stamps its Gerans with various designations that raise alarms within Ukraine. Ukrainian intelligence takes these new announcements westward in such a way as to strike fear into the hearts of their supporters, that a new generation of these aerial terrors is coming. Beyond just Geran versions 1, 2, and 3, there is an alphabet soup of sub-models that present various new threats.

The style of make and model implies a degree of serialization that is lacking in Russia’s Alabuga factory. These announcements of new models often imply that Alabuga is about to switch its manufacturing over  in a way that will make all prior operating systems obsolete, akin to the rollout of a new iPhone.

The reality is that the war’s economics demand massive swarms of Shaheds that far outstrip Russia’s ability to produce jet engines or high-end inertial guidance systems. Indeed, the “Geran-3” that Ukraine’s HUR spotlit back in September is powered by a Turbofly JT80, a Chinese-made hobbyist engine that despite the impressive term “Turbojet Engine,” offers a mere 80 kilograms of thrust.

Russian Shahed manufacturing is piecemeal, and the units even within a given “series” of higher end “Gerans” are far from standardized. They simultaneously fill their swarms with cheaper models, known as “Gerberas,” that often don’t have a warhead and feature frames made of polystyrene, among other materials. And even the Gerans with turbojet engines don’t fly top speed their whole way through Ukraine, both in order to avoid drawing the best of Ukrainian firepower to the most dangerous Shaheds and because they’ll burn through their fuel too quickly.

The practice of economical camouflage of these swarms works, and it works in large part because standard radar detection is not up to the task of distinguishing between drones, let alone between Shaheds with different internals. Like Trojan horses, these slip through the best of defenses. It is an underappreciated force multiplier that has frustrated Ukraine’s air defense enormously.

Much of the workflow of any Ukrainian anti-drone team consists of parsing out whether what they’re looking at on radar is aerial detritus or flocks of birds. Even though a Gerbera, or a dummy drone, on the ground looks nothing like a Shahed, it’s tough to make that decision while it’s in the horizon in which an anti-Shahed pilot can hope to intercept it.

Intermittent guidance but automated evasion

Shahed interception is moreover growing increasingly complicated.

Dronemakers are as captivated by autonomy as certain carmakers — and having no more luck bringing a working version to market. While the most recent numbers are not fully public, the overall Ukrainian defense budget is just over half of Tesla’s most recently reported annual revenues. The two forms of autonomy face different hurdles but are certainly comparable projects. Though, for example, road-based cars have to cope with constantly moving traffic surrounding them, the roads themselves don’t change nearly as much as the parts of a battlefield you are blowing up. And if you’re economizing on components, a one-time use kamikaze drone will never merit the kind of camera a developer would outfit on a car meant to survive.

The comparison to Tesla, which has been pouring money into self-driving vehicles since well before 2022, suggests that, despite hype, Ukraine’s drones are far from anything we’d consider autonomy. The vision of drones being able to identify targets or, world of wonders, coordinate among themselves without crashing into each other on the way down to their final strike, remains prospective. As a rule, Shaheds don’t travel in horizontal flocks so much as in punctuated lines, though Russia launches them from airfields all around its border with Ukraine.

Some Shaheds are outfitted with Starlink terminals brought to Russia via third parties for ongoing guidance through the fog of war. More feature connection via Ukrainian telecoms SIM cards and mobile internet whose supply chains are very likely also smuggled, but this time out of Ukraine and into the arms of the enemy. In either case, this is not the 5G guidance of shorter-ranged Mavics or certainly the technicolor of fiber-optic FPV drones. These are, rather, intermittent guidance aids.

Beyond AI monitoring

The flip side of artificial intelligence within the deep-strike drone sector is defense. A number of Ukrainian developer teams are working on applying artificial intelligence both to detect Shaheds and guide anti-Shahed drones to their targets.

Both tasks are in theory much easier than, say, deep-strike AI guidance. Right now a mass of Ukrainians currently sit in pick-up trucks in fields outside of major cities waiting for Shaheds every night. On any given night, an infinitesimal share of these air defense units see any action. A successful network of AI nodes that pick up audio or visual signs of Shaheds and then launch and guide interceptor drones scattered liberally across Ukraine would unlock that manpower.

But AI remains remarkably unreliable for both tasks in that anti-air killchain. The hopes placed in machine vision for air defense remain far from practical realization, with the process of shooting down a Shahed still dependent on the talent of individual remote pilots working through a control interface much like a grayscale Xbox.

There are, however, alternatives. One more traditional solution for detection is simply building radars that detect smaller flying units, like the Snov and Phoenix systems that Ukraine is developing but still is not fielding in sufficient quantities.

Another trick that may beat out machine visions is audio detection. Many of Ukraine’s air raid sirens are already rigged up with detectors that pick up sound, most notably Zvook, itself the Ukrainian word for “sound.” Audio can pick up the difference between twin-stroke and jet engines more effectively than radar can. And audio guidance in Ukrainian air-defense drones is also yielding more tangible results than machine vision.

While audio guidance can’t match the high-precision and rapid response sensors of, say, a Patriot PAC-3 missile, Shaheds do not require that degree of precision. Even the jet-powered Shaheds are slow relative to classic missiles, stay within a few kilometers of the ground and are famously noisy, with Ukrainian slang referring to them as “mopeds.” And it goes without saying, audio seeking is much much cheaper — key when you are trying to arm thousands of Shahed interceptors nationwide.

The United States military today is having a hard enough time rebooting electronic warfare units that have lain dormant since early on in the global war on terror. Maybe the most critical lesson about Shaheds is that both launching them and defending against them does not require technological marvels. Both tasks however require functional and extensive networks of competent operators, preferably full-time military personnel. Recent history suggests that the Pentagon has a much harder time delivering those kinds of teams than ambitious weaponry, which may prove the greatest stumbling block to a U.S. unmanned program of all.

Image: Shahed Drone in Sumy Oblast. 10.17. 2024. Image Credit: National Police of Ukraine, CC BY 4.0 <https://creativecommons.org/licenses/by/4.0>, npu.gov.uavia Wikimedia Commons