A professional-grade, real-time Ground Control Station (GCS) for autonomous drone fleets. This system utilizes a Data Fusion Architecture to ensure resilience, precision, and historical analysis capabilities.

The system does not rely on simple pass-through networking. Instead, it implements a Repository Pattern:

• Dual Ingestion: Merges high-frequency WebSocket Telemetry (Live) with robust REST Snapshots (Historical/Offline).
• Conflict Resolution: Smart logic ensures that stale snapshots never overwrite fresh live telemetry.
• Result: The UI never flickers or shows "No Data" when a drone disconnects. It simply transitions to a "Stale" or "Offline" state, preserving the last known location.

A built-in "Black Box" flight recorder:

• Automatic Recording: Every telemetry packet is cached in a circular buffer history.
• Visual Scrubbing: A slider allows operators to replay past flights instantly.
• Ghost Markers: Displays a "Ghost Drone" on the map to compare historical position vs. live position.
• UI Hijacking: The Replay Controller temporarily "hijacks" the Detail View to show historical battery/speed data without corrupting the live network stream.

The map system has evolved beyond static images into a dynamic Tile Rendering Engine:

• Lazy Loading: Only downloads map tiles (OSM/Satellite) for the area currently visible on screen.
• Performance Optimized: Tiles outside the viewport are unloaded to save memory.
• Dynamic Panning: Features "Logarithmic Damping"—panning is fast when zoomed in (Street View) but heavy/precise when zoomed out (Country View).

The platform solves the "Who owns what?" problem with an Active Slot Architecture:

• Slot 0 (Primary): The focused drone (Cyan theme).
• Slot 1 (Secondary): Comparison view (Orange theme).
• Visual Feedback: Map markers and UI cards automatically change color/border based on which slot controls them.

The codebase follows a strict Unidirectional Data Flow:

Network/Disk → Repository → Events → UI Controllers

• Core/: DroneStateRepository, MainThreadDispatcher.

• Network/: WebSocket client (NativeWebSocket) and REST parsers.

• Map/:

• Tiles/: Tile providers and lazy-loading renderers.

• UI/: DroneCardUI, FleetUIManager, Slot logic, Buttons and Map Related UI.

• Drone/: 3D visualizers (Propeller spin, Tilt) and Physics.

• Models/: shared C# data contracts (DroneState, DroneTelemetryData).

• Attach DroneNetworkClient to a persistent GameObject.
• Configure the WS URL (e.g., ws://localhost:5101/telemetry) and HTTP URL for snapshots.

• The MapPanelController handles zoom/pan sensitivity.
• The MapTileRenderer handles cache size and tile servers.

The system refuses to issue dangerous commands to "Ghost" drones.

• Rule: If no heartbeat is received for >5 seconds, the drone is marked STALE.
• Effect: "Arm/Takeoff" buttons become non-interactable. The Map Marker fades to 40% opacity.

• If the Replay Controller loses its link to the Detail View, it auto-detects it on the next user interaction.
• If the Network drops, the Repository retains the last known state, preventing UI errors.

1. Repository over Direct Events: We moved to a Repository so that late-joining components (like the Map) can ask "Where is everyone?" immediately on startup, rather than waiting for the next network packet.
2. Replay Override vs. State Change: We chose to "Hijack" the UI for replays rather than overwriting the actual drone data. This ensures that Live Alerts (e.g., "Battery Critical") can still trigger in the background even while the user is watching a replay.
3. Tile System: Static maps don't scale. The tile system allows the fleet to move anywhere in the world without the developer needing to manually update background images.