Project brief
A counter-UAV (C-UAV) system using software-defined radio arrays for passive radio frequency detection of drone control links and video transmission signals. The system uses AI-based signal classification to identify drone model from RF fingerprint, direction-finding antennas for bearing estimation, and a directed RF jamming module for neutralization (subject to regulatory licensing).
The expanded narrative focuses on the passive detection and analysis elements in the source record. It does not add operational deployment evidence.
The engineering challenge
Organize radio observations, classification, bearing estimates, and event records into a monitoring architecture.
Engineering approach
- Separate passive acquisition from signal classification and track estimation.
- Associate detected signal features with bearings and event history.
- Present operator alerts and system state through a supervisory interface.
Features & capabilities
- RF detection: 400MHz–6GHz wideband passive monitoring via 4-channel SDR
- AI signal classifier: CNN-based drone model recognition (94% accuracy, 50 models)
- Direction finding: 4-antenna MUSIC algorithm, ±2° bearing accuracy at 5km
- Target tracking: Kalman filter bearing + range (altitude correlation)
- Jamming module: software-selectable 2.4GHz/5.8GHz/900MHz (licensed operation)
- Alert integration: SMS, email, IP camera PTZ slew-to-target
- Automatic logging: detection events, bearing history, signal characteristics
- Mobile van-mounted and fixed installation variants
Software & engineering tools
USRP B210 SDR × 4, GNU Radio, TensorFlow signal classifier, Python backend, custom antenna array