Summary
A secured facility required a flexible, vendor-neutral infrastructure for perimeter surveillance using fixed sensors, ground robots and aerial drones.
WiseVision provided:
- WiseOS as a ROS 2-based operational layer
- MCP ROS2 for integrating AI-assisted workflows and operator tools
- A high-fidelity Digital Twin environment for training and scenario simulation
Stack Used
- WiseOS
- MCP ROS2
- Digital Twins (O3DE / Isaac Sim / Genesis)
- ROS 2
- LoRaWAN (optional)
Key Benefits
- Vendor-neutral infrastructure
- Safe AI adoption via sim-before-deploy
- Realistic training scenarios
- Full auditability and control
Challenges
Mix of existing and new assets from different vendors
Need to experiment with surveillance tactics without affecting live operations
Requirements for auditability and control over what AI agents can do
Desire to train operators and evaluate autonomous policies safely
Solution
1. WiseOS – Unified Perimeter Fabric
WiseOS connected:
- Fixed perimeter sensors (cameras, motion, RF, optional LoRaWAN) into ROS 2
- Ground and aerial platforms as ROS 2 nodes or via bridges
- Telemetry and alerts into Data Black Box (InfluxDB)
- Status and events from all assets in one place
2. MCP ROS2 – Operator & AI Tools
MCP ROS2 exposed tools with role-based controls:
- List perimeter assets
- Run patrol routes
- Get recent alerts
- Human-in-the-loop for restricted zones
- All actions logged for review
3. Digital Twin – Training & Evaluation
A Digital Twin of the facility with O3DE / Isaac Sim / Genesis:
- 3D representation of buildings, fences and terrain
- Simulated robots, drones and sensor coverage
- Same interfaces as real operations
- Training and policy testing
Architecture Overview
- Sensors, robots and drones → ROS 2 → WiseOS connectivity & telemetry
- WiseOS → InfluxDB → event logs and analysis
- MCP ROS2 → operator tools and AI-assisted workflows
- Digital Twin → training, testing and visual context
Digital Twin Use Cases
🎓 Operator Training
Training in simulated incident scenarios without risk to live operations
🧪 Policy Evaluation
Evaluation of new patrol routes and policies before deployment
🤖 AI Testing
Testing AI-driven behavior before allowing it on live assets
Outcomes
This pattern typically provides:
✅ Vendor Independence
- A neutral infrastructure layer independent of specific robot vendors
- Freedom to choose best-of-breed components
✅ Safe AI Adoption
- Sim-before-deploy workflows for AI-assisted surveillance
- Risk-free experimentation in Digital Twin
✅ Better Preparedness
- Realistic training through Digital Twin scenarios
- Improved operator readiness
✅ Full Auditability
- Complete logging of all actions and AI decisions
- Role-based access controls
Next Steps
The same architecture can be extended to larger sites, multiple facilities, or integrated with existing command-and-control tools
Note: Defence projects are evaluated individually with appropriate controls and NDAs.