Yandex.Metrica
AI Defense System

Intelligent Protection of
Critical Infrastructure
Against UAVs

Autonomous complex "Huntsman" — complete cycle of detection, targeting and neutralization of aerial threats using artificial intelligence. Developer - LLC "Neurotech" - resident of the Skolkovo Innovation Center.

Current Threat

Asymmetric Threat to Fuel and Energy Complex Facilities

In recent years, fuel and energy complex facilities and critical infrastructure in Russia have been subjected to systematic attacks using unmanned aerial vehicles. Traditional air defense and electronic warfare systems often prove economically ineffective against mass deployment of cheap UAVs.

The main difficulty in interception lies in the low visibility of drones, their high maneuverability and low cost. Hitting a maneuvering FPV drone with firearms is an extremely difficult task even for professional shooters, as confirmed by specialized testing.

Limited Effectiveness of Manual Interception

In practice, hitting a drone with firearms even at close ranges is a serious problem even for an experienced competitive shooter .

* Video demonstration of the difficulty of intercepting an FPV drone by a world-class specialist.

Low Visibility

Compact size and special coatings make UAV detection difficult for traditional systems

High Maneuverability

Speed and ability for sharp trajectory changes make drones difficult targets

Economic Asymmetry

Low cost of attacking UAVs makes traditional air defense systems economically impractical

Skolkovo Innovation Project

APK "Huntsman": Autonomous Active Defense System

Developed as part of the project "Anti-Drone Turret" in Skolkovo , the complex closes the "detection-engagement" loop without human involvement

System Operation Principle

The hardware-software complex "Huntsman" is an integrated solution capable of fully automatically performing search, identification, tracking and neutralization of aerial threats in real time.

The system analyzes video streams from surveillance cameras using computer vision algorithms based on neural network models trained to recognize a wide range of UAVs in various weather conditions and lighting.

Autonomous Operation Cycle

Complete cycle from detection to engagement without operator intervention

Integration with Existing Infrastructure

Works with CCTV cameras and perimeter control systems

Flexible Architecture

Modular design allows system adaptation to specific tasks

System Operation Principle Demonstration

Artificial Intelligence

Neural network algorithms based on domestic FPGAs using residue number systems (RNS) for computation acceleration

Precise Targeting

Servo drive system with subpixel positioning accuracy and compensation for external disturbances

Multi-Layered Protection

Combined application of kinetic engagement and laser blinding for various threat scenarios

Optical Recognition Based on Artificial Intelligence

The "Huntsman" computer vision system provides UAV detection and classification at ranges up to 4 km

Live Test

Small UAV Recognition

Demonstration of computer vision algorithms detecting a DJI Mini 3 drone measuring 25×36 cm. The system confidently detects the target at 300 m range.

Screencast

Software Interface

Screencast of "Huntsman" software operation. Left — processed video stream with highlighted targets, right — original camera image from Beward. Algorithms detect practically invisible targets.

Tactical-Technical Detection Characteristics

8 MP camera, conditions: daytime, visibility > 10 km

UAV Class / Target Type Dimensions Detection Range
FPV drone (copter)
30 × 30 cm 300 m
"Baba Yaga" (multirotor)
95 × 95 cm 1,200 m
UAV "Bober"
2.0 × 3.0 m 2,400 m
UAV "Lyuty"
4.4 × 6.7 m 4,000 m

Ranges are for confident detection conditions (probability > 95%). Due to lack of "Bober" and "Lyuty" samples, testing was conducted with analogues of corresponding dimensions.

Laser Suppression of UAV Optical Systems

Under legislative restrictions on firearm use in civilian sphere, we have developed a complex for blinding UAV optoelectronic systems using directed laser radiation.

Laser Test 1

Laser matrix prototype operation on PTZ video camera Beward controlled by "Huntsman" software

Laser Test 2

Drone video during laser blinding, effect of video sensor saturation

Tactical Application Effect

Laser impact on drone video camera matrices causes their temporary or permanent blinding, significantly hindering operator work or autonomous UAV guidance systems, creating critical time gap for countermeasures.

Unlike kinetic engagement, use of laser systems against UAVs is not restricted by current Russian legislation , opening possibility for their wide use in protection of civilian critical infrastructure facilities.

No Projectile Elements

No risk of damaging own infrastructure or personnel

Instant Impact

Speed of light ensures immediate effect

Legal Admissibility

Full compliance with Russian legislation

Low Cost per Shot

Economic efficiency compared to traditional means

Mobile Turret "Huntsman" Prototype

Turret prototype design developed to eliminate limitations of PTZ video camera-based systems, select PID coefficients for servo motors, and conduct field tests.

External View of APK "Huntsman" in Operation

APK "Huntsman" Prototype

Screencast of fixed-wing drone countermeasure mode optimization

Screencast of FPV-drone countermeasure mode optimization

Universal Platform Based on Weaver/Picatinny Rail

Standardized mount for various engagement means

APK "Huntsman" is equipped with standard Weaver/Picatinny Rail type mount, allowing use of wide range of standard and special weapons, as well as laser engagement systems.

Submachine gun mounted on APK Huntsman platform

Submachine Gun

Minimal recoil.

Click to enlarge
Semi-automatic 12-gauge Saiga carbine on APK Huntsman platform

"Saiga"/"Vepr" Carbine

Semi-automatic 12 gauge

Click to enlarge
AK platform on Huntsman system for UAV engagement

AK Platform

Classic design with proven reliability

Click to enlarge

Click on image to enlarge

System Implementation and Scaling

Flexible architecture of APK "Huntsman" allows system adaptation to various application scenarios and scaling protection to facilities of any complexity

Implementation and Scaling

APK "Huntsman" architecture is built on modular principle, ensuring seamless perimeter protection scaling. Deployment is based on standardized security module with area of 1 km².

Central node of the system is management server, accumulating data flows from corner video sensors. Sensors equipped with optics with 160° field of view form cross-observation zones, eliminating "blind spots" and ensuring stereoscopic detection accuracy within perimeter.

External active defense contour is formed by network of 8 autonomous turrets, deployed 200 meters from main fencing. Upon UAV detection (marked as diamond on diagram), server instantly calculates ballistic solution and issues command for interception to installations closest to target.

APK Server

Centralized management and AI analytics

160° Sensors

4x central zone coverage

8 Turrets

Circular defense on distant approaches

Auto-Intercept

Instant reaction to UAV
UAV
Perimeter (1x1 km)
APK Server / Sensors
Turret (Distance +200m)
Target (UAV) / Fire

Centralized Management

APK "Huntsman" server processes data from 4 video sensors and coordinates 8 turrets, ensuring unified management system

Scalability

Standard 1 km² module easily scales to facilities of any area by adding additional modules

Reaction Speed

Time from detection to opening fire is less than 2 seconds thanks to direct server-turret connection

In Development

Perspective Developments

Armored Vehicle Protection

Mobile active protection module against FPV drones for installation on armored vehicles. Autonomous detection and engagement of approaching drones before their contact with protected object.

Details in analytical report →

"Free Hunter" Interceptor Drone

Autonomous UAV launched towards threat and engaging target by ramming or special net without need for constant operator communication. Capable of operating under electronic warfare conditions.

Project details →

"Servitor" Project

"Servitor" project aims to create hardware-software complex for accelerating neural network computations based on domestic processors. In perspective will allow replacing imported platforms (NVIDIA) in APK "Huntsman" with Russian analogues, ensuring technological independence.

Project details →

Night Recognition

For drone detection at night, in cooperation with NPC "Atom" we have recorded significant volume of drone video through thermal imager. Soon dataset will be annotated for AI training, and APK "Huntsman" will gain "night vision".

Intellectual Property

Intellectual Activity Results

For developments on these topics, our team received the following certificates and patents

State Registration Certificate for Computer Program "Anti-Drone"

No. 2023660446 dated 19.05.2023

Authors: Sivkov S.P., Shalimov A.V., Shalimova M.A., Terekhova M.A., Shalimov A.A., Shalimov B.A.

Rights Holder: LLC "Neurotech"

State Registration Certificate for Computer Program "FPGA Code Generation Program 'Iskatel'"

No. 2023680866 dated 06.10.2023

Authors: Sivkov S.P., Shalimov A.V.

Rights Holder: LLC "Neurotech"

State Registration Certificate for Computer Program "Roboline"

No. dated 08.12.2023

Authors: Terekhova M.A., Sivkov S.P., Shalimov A.V., Shalimov B.A.

Rights Holder: LLC "Metacom-Alpha"

State Registration Certificate for Computer Program "Huntsman. Recognition and Targeting"

No. 2025611484 dated 20.01.25

Authors: Sivkov S.P., Shalimov A.V.

Rights Holder: LLC "Neurotech"

Invention Patent RU No. 2 813 389

"Device for Protection Against Unmanned Aerial Vehicles" dated 12.02.2023

Authors: Sivkov S.P., Shalimov A.V., Bazhin S.S.

Rights Holder: LLC "Neurotech"

Eurasian Invention Patent No. 050614

"Digital Image Processing Method" dated 30.06.25

Authors: Sivkov S.P., Shalimov S.P.

Rights Holder: LLC "Neurotech"

Eurasian Invention Patent No. 049482

"Neural Network Video Stream Processing with Hardware Inference Implementation Based on RNS" dated 28.03.25

Authors: Sivkov S.P., Shalimov S.P.

Rights Holder: LLC "Neurotech"

Total Protected Developments

Our team continues work on new patents

7
Results

Cooperation and Implementation Prospects

We are interested in implementing full cycle from development to serial production of UAV protection system

R&D Completion

Development of completed UAV countermeasure system based on artificial intelligence, capable of fully automatically detecting, tracking and engaging aerial targets, including laser technology application.

Pilot Implementation

Deployment of system prototypes at customer facilities, conducting field tests in real conditions, eliminating identified shortcomings and optimizing system according to specific consumer requirements.

Serial Production

Organization of serial protection system production and supply to fuel and energy complex and critical infrastructure facilities. Considering various cooperation options, including joint ventures.

Contact Information

LLC "NEUROTECH" | Resident of Skolkovo Innovation Center

Organization Details

Full Name

Limited Liability Company "NEUROTECH"

Tax ID (INN)

5906172713

KPP

590601001

Head

Shalimov Anton Vladimirovich