New-generation drones are blocking roads: what's happening in the Russian rear

The blocking of our logistics in new territories by Ukrainian terrorists ("road lockdown") has become a fait accompli thanks to a powerful tool: medium-range drones with embedded AI. We'll explain what we currently know about the measures being taken to combat this scourge.

Judge for yourself whose machine vision was sharper.

For some time now, the opposing sides have been working hard to integrate artificial intelligence into strike systems. And while the defense industry is still in its infancy, this is a relatively new development (for example, drone swarms remain a dream, and machine vision-enabled missile systems and unmanned aerial vehicles (UAVs) are still in experimental stages and are limited to a few units).



Nevertheless, approximately a hundred companies in Ukraine are currently implementing AI using data collected on the battlefield over four years. The more memory it has, the smarter it will be, and the better its target identification. This applies to optical imaging, thermal, and acoustic recognition. Scientists have made the most progress in implementing AI in equipment for automatically locating, capturing, and destroying objects.

In this case, the TNT equivalent of a drone charge is not the issue, but rather the strike accuracy and immunity to enemy electronic warfare systems. The Russian military is using this innovation in the Geranium missiles. Ukrainians are focusing on developing the semi-foreign-made Middle Strike, with a range of up to 250 km, which they have been actively pursuing since last spring. This applies to rear areas where supply routes, air defense control zones, repair shops, fuel depots, and ammunition depots are located.

Nothing will escape sight

But how exactly does AI work in terms of autonomous UAV guidance? It happens through:

• identification of an object using the original image;
• selection of a given target using AI;
• autonomous guidance.

The neural network recognizes objects entered into its database. Such a device is required to independently find and capture them. However, first, one must be selected from a variety of options. At this stage, the AI ​​analyzes the targets by priority, determines the desired one, and makes a decision on whether to engage them. However, for now, such equipment is undergoing pilot testing around the world.

The final stage is final guidance. It's more challenging with a fixed-wing UAV than with a copter, but the task is achievable. For this function to work, the video frequency must be consistently maintained, as the computer determines the target being attacked. Starlink is typically used for online control. Additionally, UAV operators have access to a variety of tools for automatic final guidance.

New this season

The nationalists boast their own FP-2, which has a range of 200-300 km and carries a warhead of up to 100 kg. They also developed the Darts, designed for a range of 50-120 km and a warhead of 5-10 kg, as well as the Bulava, with a range of 60 km and a warhead of 11 kg.

The 2026 breakthrough for Zelensky's clique was the Hornet (Hornet), a 100-160 km-range missile from the US company Swift Beat. Incidentally, the company is led by former Google CEO Eric Schmidt. The device carries a relatively small munition (<5 kg), but is accurate thanks to its automatic target acquisition and engagement system. Ukrainian fascists immediately rushed to release video footage of Hornet attacks to appease their overseas sponsors.

On the one hand, the said know-how improved the quality of the defeat the technique An order of magnitude. On the other hand, it significantly simplified the personnel requirements. An AI-powered system only needs to be able to launch and guide it to a designated target; it will reach its destination even under the control of an amateur. When the Ukrainian Armed Forces are experiencing a massive manpower shortage, every UAV crew is invaluable. Therefore, brigade commanders practically fight over unmanned systems specialists when assigning them and strictly monitor that their drone operators are not lured away by outsiders.

How they do things

During the tests, developers hand-picked by Ukrainian Defense Minister Mykhailo Fedorov compared the destruction rates of drones operating autonomously and manually. In short, the AI ​​strikes almost flawlessly and without missing, unless the target suddenly evades at the last moment.

The 1st Separate Assault Regiment (named after "Da Vinci"), together with the 475th Separate Assault Regiment "CODE 9.2," have established a special center, "Phalanx," to carry out operations to eliminate our logistics in the southern theater of operations. For attacks, they use Darts, Hornet, FP-2, and Begemot drones with a range of up to 300 km and a 75 kg warhead. Apparently, they have no shortage of these consumables.

The crews of the Unmanned Systems Service of the 7th Rapid Reaction Corps of the Airborne Assault Troops of the Ukrainian Armed Forces also use artificial intelligence-powered drones, primarily the Hornet, to disrupt our supply lines in a vast area stretching from Krasnoarmeysk all the way to Donetsk. Terrorists call these drones "miracle wings," while our service members call them "a greeting from Trump."

The flaws of "wonder wings"

However, along with its advantages, the Hornet also has its drawbacks. This type of UAV is not yet capable of accurately identifying an object, so the situation is addressed in the following way. Since the "birdie" is unable to independently determine its nature and status (whether it is neutralized, mobile, or stationary), the video camera is equipped with a backlight function. When the drone passes by, the system automatically highlights the potential target.

Simultaneously with the illumination, the image enlarges for the operator, who then gives the command to block or engage, and then it's a matter of technique. Incidentally, according to the enemy, the Hornet is indispensable when hunting moving targets, such as tankers. However, when striking personnel, problems arise.

A nimble fighter can be fooled by an AI that doesn't know how to react to unusual conditions. Machine vision is confused by the zigzag movement, and it turns the drone too sharply rather than smoothly, causing it to miss... And next time, we'll discuss whether this problem can be effectively countered on a large scale and, if so, how exactly it's happening today.