Detect
Ground and distributed sensors establish a verified picture of low-altitude activity.
Autonomous counter-UAS research programme
SEE THE THREAT.
PREDICT THE MOVE.
MATHryoshka turns uncertain drone behaviour into a physically bounded, continuously updated response problem.
AIR PICTURELIVEDECISION ENGINEACTIVERESPONSE LOOP20 Hz
RIGA, LATVIARESEARCH → ENGINEERING → VALIDATION SYSTEM CONCEPT ACTIVE
01 / THE MISSION
A PROTECTIVE DOME FOR
THE LOW-ALTITUDE DOMAIN.
AutoFly is developing an integrated response chain for uncooperative aerial threats: from detection and classification to prediction, autonomous guidance and controlled interception.
One system. One continuously updated air picture.
Designed for critical infrastructure, restricted zones and defence validation environments.
02 / SYSTEM ARCHITECTURE
FROM SIGNAL
TO RESPONSE.
Each layer reduces uncertainty before the next decision is made.
Understand
Multi-sensor fusion classifies the object, estimates its state and measures confidence.
Predict
MATHryoshka calculates the target's physically reachable space under uncertainty.
Respond
The system selects a controlled response and continuously replans the interceptor path.
03 / SCIENTIFIC CORE
MATHRYOSHKA
A reachable-space model for an unknown aircraft.
Developed in Dr Ivan Panov's PhD research at Aarhus University, the method estimates where an uncooperative drone of an unknown model can physically move within a defined time horizon.
Dms = s(T) + dimpact
- 01 Target type and observed velocity
- 02 Aerodynamic and performance boundaries
- 03 Worst-case reachable-space calculation
- 04 Continuous trajectory replanning
04 / AIRBORNE RESPONSE
BUILT TO CLOSE
THE DISTANCE.
A modular fixed-wing development platform connects the mathematical prediction layer to a real, measurable flight response.
VECTOR MODELACTIVEGUIDANCE LOOP20 Hz
TARGET TRACK / CONTINUOUS UPDATE
PHYSICAL DEVELOPMENT PLATFORM
THE INTERCEPTOR
ALREADY EXISTS.
The carbon-fibre development prototype provides a physical platform for propulsion, flight-control, sensor and guidance integration.
AI-assisted flight controlModular mission payloadGround & flight validation
05 / DEVELOPMENT PATH
FROM MODEL
TO FIELD EVIDENCE.
The programme advances one measurable risk at a time.
TRL 3
Scientific foundation
Mathematical modelling, simulation and an existing flying platform.
CURRENT FOUNDATIONTRL 4
Integrated demonstrator
Sensor inputs, autonomy stack and closed-loop guidance validated together.
INTEGRATION GATETRL 5
Relevant-environment proof
Repeatable field trials with evidence, logs, metrics and failure analysis.
PROGRAMME TARGET06 / PEOPLE & COLLABORATION
SCIENCE MEETS
FLIGHT ENGINEERING.
AutoFly AI Robotics brings together UAV development, autonomous systems research and field validation around one integrated mission.
Deniss Isakovs
AutoFly AI Robotics. Product direction, interceptor development and partnership programme.
Dr Ivan Panov
PhD research in autonomous UAV operations, safe separation with uncooperative aircraft and 4-D trajectory planning.
Collaboration pathwayAcademic validation · Defence end-users · Test environments · Engineering partners
07 / CONTACT
LET'S BUILD THE
NEXT LAYER OF DEFENCE.
We are opening conversations with research, validation and defence-technology partners.
PARTNERSHIP ENQUIRIES
Riga, Latvia · European Union