ONYX GRID
Passive Drone Detection Mesh Network
[VISION] — Target Deployment Architecture
Onyx Grid is a decentralized infrastructure of smart sensor nodes designed to detect, track, and alert against unauthorized UAV activity in real-time. The fully deployed system operates as a passive, mesh-networked detection layer requiring no external cloud connectivity.
System Architecture
- ▸ Distributed Sensor Nodes: Multi-modal detection (Video, Audio, RF/SDR) deployed across perimeter or airspace region.
- ▸ Mesh Networking: LoRa/LTE-based peer-to-peer coordination enabling fully decentralized, air-gapped operation.
- ▸ C2 Gateway: Local Java/Quarkus gateway aggregating MQTT telemetry, persisting to SQLite, and broadcasting CoT (Cursor-on-Target) via UDP.
- ▸ Tactical Console: Dark-themed web UI (MapLibre GL JS) with real-time node markers, threat vectors, and offline-capable operations.
Core Capabilities
Emits no signals; undetectable to adversary RF analysis.
Sub-second local processing; zero cloud dependency.
Survives cellular/WiFi jamming; distributed redundancy.
Standard CoT/ATAK multicast; works with existing C2 systems.
[ENGINEERING PROGRESS] — Current State (TRL 4)
SENSOR NODE SUBSYSTEM
ACTIVECurrent Status: Component validation framework established. Multi-sensor integration pipeline configured for video and audio analysis. Edge computing architecture eliminates traditional bottlenecks in real-time signal processing.
Implemented Features
- ✓ Multi-Sensor Framework: Extensible architecture for video, audio, and future RF detection.
- ✓ Edge Deployment Target: Raspberry Pi 5 + Hailo-8L (13 TOPS INT8) accelerator.
- ✓ Telemetry Isolation: MQTT bridge for safe external communication without exposing internal mesh.
- ✓ Lab Validation: Architecture tested in laboratory environment with controlled sensor inputs.
Hardware & Sensors
Development Roadmap
- [-] RF Detection Integration: Expand to radio frequency domain analysis for comprehensive drone signature detection.
- [-] Custom AI Models: Develop and optimize machine learning models tailored to Hailo-8L constraints.
- [-] Hardware Engineering: Enclosure design, passive cooling, environmental hardening for field deployment.
C2 GATEWAY — TELEMETRY AGGREGATION
OPERATIONALCurrent Status: Fully functional MVP. Ingests sensor telemetry over mTLS MQTT, persists state to PostgreSQL, broadcasts tactical network information via CoT multicast (239.2.3.1:6969) and WebSocket streaming. Ready for configuration abstraction and cloud staging.
Core Subsystems
Mutual TLS (mTLS, TLS 1.3, port 8883) with Eclipse Mosquitto 2.0. Consumes node heartbeats and threat alerts; validates and clamps payloads (speed 0–150 km/h, azimuth 0–359°).
PostgreSQL 16 with Flyway migrations. Tracks node identity, geolocation, lifecycle state, threat events (classification, azimuth, confidence, speed), and geofenced perimeters.
Cursor-on-Target (CoT) XML multicast for ATAK integration. Real-time WebSocket streaming of node state and threat detections to authenticated console clients (JSON).
MapLibre GL JS with Protomaps PMTiles (offline-capable vector tiles). Real-time node markers (color-coded: green=ALIVE, gray=OFFLINE, red=THREAT). Directional threat arrows. Geofenced perimeter visualization.
Console UI — Tactical Screenshots
Security & Authentication
- ✓ OIDC Authentication: Web console protected via Keycloak (Authorization Code + PKCE flow).
- ✓ HTTPS: Self-signed for dev/test; production CA-signed certificate support.
- ✓ Embedded PKI: Local certificate authority for mTLS node commissioning (no external infrastructure required).
Development Roadmap
- [-] Dynamic Configuration Service: Replace hardcoded parameters (multicast group/port, GC intervals, thresholds) with runtime REST API.
- [-] Cloud Infrastructure Staging: Deploy to AWS/Azure with observability (structured logging, Micrometer metrics, APM).
- [-] Live Demo with Real Sensors: Commission field-deployed nodes; transition simulator to shadow mode for validation.
Technology Stack
[ROADMAP] — Immediate Technical Priorities
Near-Term Sprints (TRL 4 → TRL 5)
-
[ ]
RF Detection Integration
Expand detection capabilities to radio frequency domain. Integrate SDR (Software Defined Radio) analysis into active MQTT telemetry bus for comprehensive drone signature coverage.
-
[ ]
Decentralized Mesh Networking
Enable peer-to-peer coordination between sensor nodes. Implement LoRa mesh for fully air-gapped, decentralized deployment. Eliminate single-point failure via redundant routing.
-
[ ]
Custom AI Model Optimization
Develop and validate machine learning models for acoustic and visual drone signatures. Implement INT8 quantization targeting 13 TOPS Hailo-8L accelerator. Achieve <100ms inference latency on edge.
-
[ ]
Physical Hardware Engineering
Design robust node enclosures with passive cooling and thermal management. Implement weatherproofing (IP67 minimum) for outdoor field deployment. Validate mechanical resilience under operational stress.
-
[ ]
C2 Configuration Service (Stage 13)
Abstract hardcoded parameters into runtime REST API. Enable operators to tune multicast groups, GC intervals, and node thresholds without redeployment.
Medium-Term Goals
- [-] Cloud Infrastructure Staging (Stage 14): Migrate to AWS/Azure with containerized deployment. Establish public staging URL. Integrate observability (structured logging, Micrometer metrics, optional APM).
- [-] Live Demo with Real Sensors (Stage 15): Commission actual field-deployed Onyx Grid nodes. Transition simulator to shadow mode. Establish persistent live demo for stakeholder evaluation and operational training.
Production Readiness Criteria
System is feasible for small-scale tactical mesh networks (50–200 edge nodes per gateway) upon completion of:
- • Real sensor integration (end of RF/Mesh/AI pipeline)
- • Field-hardened hardware design and validation
- • Dynamic configuration and cloud staging infrastructure
- • End-to-end operational testing with distributed deployment scenario