Skip to main content

Components

The FleetCore ecosystem consists of three primary active components that work together to manage drone fleets.

1. Backend Server (Java/Quarkus)​

Central coordination server. Manages persistent state and configuration for the entire fleet, organized by Outposts.

Key Responsibilities:​

  • Outpost Management: High-level logical organization of drones and groups into geographic outposts.
  • Drone Onboarding: Registers new drones, provisions AWS IoT Certificates and Kinesis Video Signaling Channels.
  • Mission Orchestration: Uses FleetCoreLib to generate mission bundles (survey area partitioning or manual waypoints) and dispatches them via AWS IoT Jobs.
  • Job Scheduling: Supports immediate or scheduled mission execution.
  • Maintenance Tracking: Keeps records of drone maintenance history and pending tasks.
  • Data Persistence: Stores telemetry, detection events, and mission history in the database.

Tech Stack:​

Database Schema Design​

The backend manages a PostGIS-enabled PostgreSQL database. The schema is designed to handle geospatial drone data, mission logs, fleet maintenance, and other required records.

Database Schema Diagram

2. Desktop Client (Tauri/React)​

The Ground Control Station (GCS) used by operators to plan missions, manage fleets, and control drones in real-time.

Key Responsibilities:​

  • Outpost Management: Interface for managing multiple geographic regions and the groups assigned to them.
  • Fleet Maintenance: Dashboard for tracking drone health status and completing maintenance logs.
  • Mission Planning: Interface for defining survey boundaries (group missions) or manual waypoints (solo missions).
  • Real-time Monitoring: Low-latency WebRTC video feed streamed directly from the active drone.
  • Manual Control: Integrated support for gamepads and touch/mouse controls with real-time subsystem health checks.

Desktop Home Page​

Desktop Home Page Figure 1: Landing page of the client showing active outposts.

Desktop GCS Overview​

Desktop GCS Planning UI Figure 2: Mission planning screen for area coverage.

Low-Latency Manual Control​

WebRTC & Manual Control UI Placeholder
Figure 3: [Screenshot W.I.P] Manual drone control interface with telemetry and subsystems.

Tech Stack:​

3. Onboard Agent (Python)​

The coordinator software of the drone, running on it's companion computer (Raspberry Pi 5).

Key Responsibilities:​

  • Flight Controller Bridge: Communicates with the PX4 flight stack via MAVSDK.
  • Job Consumer: Listens for AWS IoT Jobs, downloads mission bundles from S3, and executes them autonomously.
  • State Management: Uses a dedicated State Machine to track the drone's lifecycle.
  • Telemetry Provider: Publishes real-time GPS, battery, health, and AI detections to AWS IoT Core.
  • Video Streamer: Manages GStreamer pipelines to encode camera data and stream it via WebRTC.
  • Manual Controller: Handles low-latency control commands from FleetCoreDesktop via WebRTC data channels.

Onboard Agent In Simulator​

Onboard Agent Simulator Figure 4: The drone agent controlling a simulated drone in Gazebo Classic.

Tech Stack:​

  • Language: Python 3.13+
  • Drone Link: MAVSDK, MAVLink
  • Streaming: GStreamer + aiortc
  • Messaging: AWS IoT SDK v2