This repository implements an autonomous mission control system for scanning, detecting, and localizing Bluetooth Low Energy (BLE) tags using the PX4 flight stack and MAVSDK.

This system provides a foundation for:

• Search and Rescue (SAR): Locating beacons or missing persons carrying BLE devices.
• Asset Tracking: Autonomous inventory scanning in large warehouses or outdoor yards.
• Algorithm Development: A safe, simulation-first environment for testing complex path planning and localization logic.

This project uses a split-repository approach:

• PX4-Autopilot/ (Submodule): Points to the official PX4-Autopilot repository. It contains the firmware and simulation environment.
• mavlink_scripts/: Contains the custom Python code for the autonomous mission.
  • core/: Main mission logic (intelligent_mission.py).
  • localization/: Particle filter and Gaussian Process mapping.
  • simulation/: Bluetooth tag simulation and environment setup.
  • interface/: Visualization and dashboard tools.
• scan-dashboard/: Web-based dashboard for mission monitoring (WIP).

Since this repo uses submodules, clone it recursively:

git clone --recursive https://github.com/your-username/px4_mavsdk_ble_scan.git
cd px4_mavsdk_ble_scan

If you already cloned it without --recursive, run: git submodule update --init --recursive

• PX4 Toolchain: Follow the PX4 Dev Guide to install toolchains for your OS.
• Python 3.8+: Required for MAVSDK scripts.
• MAVSDK: Install the Python library:

  pip install mavsdk numpy matplotlib websockets pyyaml

Open a terminal and launch the PX4 simulation:

cd PX4-Autopilot
make px4_sitl gz_500

Open a new terminal and run the tag spawning script:

python3 mavlink_scripts/simulation/spawn_tag_markers.py

Open a new terminal and run main mission script:

python3 mavlink_scripts/core/intelligent_mission.py

The system implements an "Intelligent Mission" controller that:

• Predictive Exploration: Uses a Gaussian Process (GP) to model signal uncertainty and guide the drone to areas with the highest probability of tag discovery.
• Precision Localization: A Particle Filter integrates noisy RSSI (signal strength) data to converge on the exact coordinates of static BLE tags.
• Simulation-First Design: Full integration with Gazebo (GZ) via the PX4 SITL stack.

graph TD
    subgraph Onboard_Compute [Onboard Computer]
        Mission[Mission Controller] -->|Waypoints/Velocity| MAVSDK
        MAVSDK <-->|Mavlink| PX4[PX4 Autopilot]
        
        Mission -->|Query| BLE_Sim[BLE Simulator]
        BLE_Sim -->|RSSI| Mission
        
        Mission -->|Observations| Planner[Gaussian Process Planner]
        Planner -->|Next Waypoint| Mission
        
        Mission -->|RSSI + Pos| PF[Particle Filter]
        PF -->|Tag Location Est| Mission
        
        Mission -->|State| Dashboard[Web Dashboard]
    end
    
    subgraph Simulation
        PX4 <--> Simulator[Gazebo / JMavSim]
    end

Gaussian Process Belief Map

Particle Filter Localization

• PX4 Updates: The PX4-Autopilot folder is a submodule. It points to a specific commit of the upstream repo. To update it:

  cd PX4-Autopilot
  git checkout main
  git pull
  cd ..
  git add PX4-Autopilot
  git commit -m "Update PX4 submodule"