Software-Defined Drone Detection

The counter-drone market is worth $4.9 billion and projected to reach $36 billion by 2035. Current solutions from DroneShield, Dedrone, and others cost $100,000 to $500,000+ per installation.

The hardware is a commodity. The intelligence is not.

The current landscape

Every major counter-drone system works the same way: RF sensors detect drone control signals, radar tracks the airframe, acoustic microphones listen for propeller signatures, and cameras provide visual confirmation. The sensor fusion and classification happen in proprietary software.

The hardware components are increasingly commoditized:

• Software-defined radios that cover 1 MHz to 6 GHz cost $300 (HackRF One)
• MEMS microphone arrays cost $5-50 per element
• The US Department of Defense has already published open-source drone detection code (porglet on GitHub)

What makes DroneShield and Dedrone valuable isn’t their sensors. It’s their ML models and protocol databases covering 300+ drone models. That’s the moat. And it’s a moat that shrinks every time someone publishes a new open-source implementation.

What software-defined means

Instead of purpose-built hardware, software-defined drone detection uses commodity radio hardware running intelligent software:

1. Spectrum monitoring: A $300 SDR continuously scans 1 MHz to 6 GHz
2. Baseline learning: AI learns the normal RF environment at each location
3. Anomaly detection: When a new signal appears, it’s analyzed against known drone protocols
4. Classification: ML models identify the drone type (DJI, FPV racing, custom)
5. Alerting: Real-time notifications with confidence scores

The key insight: every drone emits RF signals (control links, video downlinks, Remote ID beacons). These signals have distinctive characteristics that ML can learn.

Why this matters for defence

Ukraine’s experience proved that counter-drone capability is no longer optional for any military force. The proliferation of cheap commercial drones means that $500 drones are being used as weapons. Defending against them with $100K+ sensor systems doesn’t scale.

Software-defined detection at $300 per node with $500-2,000/month for the intelligence layer makes it possible to deploy counter-drone awareness at every forward position, every embassy, every critical facility.

The open-source foundation

The DoD’s porglet project demonstrates that the technical approach works. Open-source DJI DroneID demodulators prove that consumer drones can be detected and identified by SDR. The ML training data is becoming available.

What’s missing is a productized, accessible, subscription-based service that makes this capability available to facilities that can’t afford a $500K DroneShield installation.

What we’re exploring

At Symvek, we’re researching software-defined counter-surveillance technology that includes drone detection as one layer of a broader awareness platform. Our Sentinel surveillance mapping technology (498,500+ cameras across 183 countries) demonstrates our capability in surveillance infrastructure intelligence. Extending this to airspace awareness is a natural evolution.

Counter-drone detection is fundamentally a signal classification problem. The same SDR hardware and ML approaches that detect surveillance bugs can detect drone control links. The platform is the same.