How Ukraine uses AI to guide long-range drone strikes through electronic warfare and deep into Russian-controlled rear areas

Ukraine’s Ministry of Defense (MoD) says artificial intelligence is increasingly integrated into so-called “middle strike” systems, enabling long-range drones to operate in heavily contested electronic warfare environments and strike targets at operational depth.

Officials say the systems are part of a broader program aimed at disrupting Russian logistics and supply chains, with Ukraine reporting a sharp increase in strikes on transport routes, depots, command posts, and equipment concentrations in recent months.

The MoD explains that the use of AI is intended to overcome two key constraints in this operational zone: heavy electronic warfare interference that disrupts satellite navigation, and dense air defense coverage that limits predictable flight paths.

AI takes over control at target approach phase

According to the ministry, AI systems are not active throughout the entire mission. Drones are initially flown under operator control, but once they reach a designated target grid, onboard systems take over navigation and targeting.

At that stage, onboard computers and optical sensors begin real-time analysis of the surrounding area, switching from manual guidance to autonomous control for the final approach.

Visual navigation without GPS using terrain matching

The ministry says drones equipped with AI systems are able to operate without reliable satellite navigation by using onboard cameras and computer vision models.

These systems continuously scan terrain features such as roads, rivers, and landscape contours, and compare them against preloaded high-resolution satellite imagery. 

This allows the drone to determine its position based on visual correlation rather than GPS signals, which are often degraded or spoofed by electronic warfare.

Automated target recognition and terminal strike adjustment

In the terminal phase, AI systems analyze live video feeds to identify military equipment based on trained visual patterns. The system is designed to distinguish between different types of targets and automatically assign a tracking marker once a valid target is recognized.

Once a target is locked, the AI adjusts flight controls in real time to guide the drone into a final dive trajectory.

The ministry also says the system can reduce misidentification risks by distinguishing between real equipment and decoys using multiple indicators, including geometry, surface texture, and thermal signatures.

Russian forces are increasingly trying to counter these AI-assisted drone targeting systems using methods such as painting military vehicles with high-contrast zebra-like stripes. Experts say the markings are designed to confuse computer vision models trained to identify standard vehicle shapes and camouflage patterns.

Route planning shaped by electronic warfare conditions

The MoD says AI is also used before launch to optimize flight paths. Planning systems incorporate intelligence on Russian air defense and electronic warfare deployments, selecting routes that minimize exposure.

This includes analysis of radar coverage, identification of terrain masking opportunities, and the use of blind spots in air defense systems. The goal, according to officials, is to ensure drones can reach operational depth targets despite dense electronic warfare and interception layers.

“Logistics Lockdown” program and expanded strike depth

Ukraine says these systems are being deployed as part of a wider “Logistics Lockdown” program aimed at increasing pressure on Russian military supply chains.

Officials say “middle strike” operations are now being scaled to target Russian logistics, infrastructure, and military assets at distances of up to 200 kilometers behind the front line.

The MoD says the approach is designed to combine intelligence, long-range strike capability, and automation to increase the tempo and effectiveness of attacks on Russia’s rear support systems.