SILENT DOME
Technology

Four sensors.
One unified threat picture.

A deep dive into the layers, the AI Core, and the engineering principles behind Silent Dome.

03Sensor Fusion · Illustrative Simulation

The Fusion Engine.
Four sensors, one corroborated track.

The Silent Dome fusion engine correlates acoustic harmonics, RF protocol fingerprints, radar tracks, and EO/IR classification — using a Bayesian per-sensor likelihood model — to produce a single corroborated track per target with a full auditable evidence chain. Numbers shown below are illustrative; real deployments tune ranges, thresholds, and integrations per site.

01
Acoustic
Audio signature detected
Rotor harmonic match · 78%
02
RF Intel
Control link fingerprinted
DJI · 2.4 GHz · operator 940m NE
03
Radar
Track established
Range 820m · alt 85m · 18 m/s
04
Optical
Visual lock acquired
Quadcopter · 92% confidence
05
Alert
Threat dispatched
Owner + monitoring + log
Silent Dome · Site Alpha

Illustrative simulation. Real deployments tune ranges, thresholds, and integrations per site.

02The Detection Stack

Each layer compensates for the others.
Together, they're decisive.

01
Acoustic Detection Perimeter
80–250 m per node
Outdoor weatherproof microphone arrays sample environmental audio continuously, running on-device AI inference to match incoming sound against a library of drone acoustic signatures. Beamforming gives bearing without a directional antenna. The acoustic layer is the cheapest, lowest-power tier — and often the first to detect a drone closing on the perimeter, especially small autonomous quadcopters running silent on RF.
  • Beamforming microphone arrays
  • AI signature library (multi-rotor, fixed-wing, FPV)
  • Direction estimation per node
  • IP65 weatherproof housing
  • Low-power edge inference (no cloud round-trip)
  • Mesh-aware fusion across multiple nodes
02
RF Drone Intelligence
Passive · long-range
Passive RF spectrum analysis listens for the command and telemetry links that drones use — control protocols, FPV video, Wi-Fi UAVs, and Remote ID broadcasts — and matches them against a constantly-updated drone protocol library. Because it's passive, it never transmits. There's nothing to license and nothing to intercept. RF Intelligence also enables operator location estimation — the human flying the drone, not just the drone itself — when the operator's controller is actively transmitting. RF is blind to autonomous drones with Remote ID disabled; fusion with radar and acoustic covers that gap.
  • Drone protocol fingerprinting (DJI, Autel, Skydio, custom)
  • Remote ID broadcast capture
  • Pilot location triangulation (when controller transmitting)
  • FPV video link detection
  • Fully passive — no transmissions
  • Continuous library updates
03
Compact Drone Radar
1–2 km
A compact, low-power radar tracks small aerial targets across the protected airspace. AI-assisted micro-Doppler analysis distinguishes drone rotors from birds, swaying trees, and other clutter. The radar continues to operate through weather and darkness — conditions that defeat optical-only systems. Multi-target tracking maintains simultaneous tracks across several incoming objects.
  • 1–2 km detection range
  • 360° or 120° sector coverage
  • AI micro-Doppler classification
  • Multi-target simultaneous tracking
  • All-weather, day and night
  • Low transmit power, FCC-compliant
04
AI Optical & Thermal Tracking
Autonomous PTZ · evidence-grade
Once acoustic, RF, or radar establishes a track, the AI Core slews a PTZ camera onto the target, locks, and begins recording. Optical zoom and (optional) thermal imaging produce visual confirmation and admissible evidence. The classifier runs on the live video stream — confirming the object class and recording the encrypted, timestamped evidence chain that fusion needs to reach a final decision.
  • Autonomous PTZ slew + lock-on
  • Optical zoom (up to 30×)
  • Optional thermal imaging fusion
  • AI object classification on-stream
  • Encrypted evidence recording
  • Frame-accurate timeline for forensics
03Engineering principles

What we optimize for.

Fusion-first

No single sensor is trusted. Every decision requires corroboration across at least two independent layers — collapsing false positives.

Edge AI

All real-time inference runs on the edge. Cloud is for fleet management and model updates, not the decision loop.

Passive by default

Detection layers are designed to be electromagnetically quiet. We don't emit unless explicitly authorized to.

Auditable

Every alert is logged with a full evidence chain: which sensors fired, what classifier, what confidence — readable in seconds.

Modular

Each layer is independent. Sites can deploy any subset. Upgrades and retrofits are first-class.

Software-defined

Sensor capability evolves through model updates and firmware, not hardware swaps. New drone protocols arrive as updates.

Silent Dome · See it before it reaches you.

Airspace is the
next perimeter.

Talk to our team about a site assessment, a tailored configuration, or a pilot deployment.

Request DemoEnterprise Inquiry