MAVSDK Drone Show (MDS) – All-in-One Drone-Show & Smart Swarm Framework

MDS is a unified platform for PX4-based drone performances and intelligent swarm missions. Whether you want to run a pre-planned, decentralized drone show using SkyBrush outputs or orchestrate a live, collaborative swarm with leader–follower clustering (Kalman-filter enhanced, dynamic role changes, leader failover), MDS has you covered.

Overview

MDS 3 combines three core components into a single, cohesive package:

1. Drone Side

   • Runs on any Linux-based autopilot platform (e.g., Raspberry Pi, NVIDIA Jetson, or similar).
   • Handles MAVSDK integration, local trajectory execution, failsafe monitoring, and dynamic formation logic.

2. Cloud Side (Optional Backend)

   • Hosts the formation-planning engine, mission dispatcher, and WebSocket/MAVLink router.
   • Provides global setpoint management, three-way startup handshake, and health-check services.
   • Can be deployed on any cloud VM (Ubuntu 22.04 / 23.04 recommended) or on-premises server.

3. Frontend (React Dashboard)

   • Full-featured React GUI for real-time monitoring and control.
   • Upload offline “ShowMode” trajectories (CSV/JSON from SkyBrush), visualize live positions, assign leaders/followers, and trigger mission modes.
   • Supports both Drone-Show mode (offline choreography) and Smart-Swarm mode (live clustering with Kalman filters, dynamic formation reshaping, and leader-failover).

In short, MDS 3 is one package that can be used for:

• Offline Drone Shows: Pre-planned, synchronized formations based on SkyBrush CSV outputs.
• Smart Swarm Missions: Cluster-based, fully decentralized leader–follower missions with robust failsafe handling, automatic leader re-election, and dynamic reconfiguration.

---

Demo Videos

• V2 Drone Show & SITL Setup
  100-Drone SITL Test in Clustered Cloud Servers | MDS Mavsdk Drone Show Version 2
  

• Smart Swarm Mode Demo
  

---

Key Features

• All-in-One Architecture

  • Shared Docker image and codebase for both offline shows and live swarm missions.
  • Single command-line interface and unified React dashboard.

• Offline Drone-Show Mode

  • Converts SkyBrush CSV/JSON into MAVSDK-compatible “ShowMode” files.
  • Global setpoint propagation ensures perfect synchronization across 10–100+ SITL instances or real drones.
  • Preflight sanity checks (battery, GPS lock, ESC health) before arming.

• Smart Swarm Mode (Live, Decentralized)

  • Clustered leader–follower architecture with Kalman-filter–enhanced state estimation.
  • Automatic leader failure detection and re-election based on configurable metrics (battery level, signal strength, drone ID).
  • Dynamic formation reshaping and per-drone role changes on-the-fly.
  • In-flight failsafe monitors communication timeouts, altimeter discrepancies, ESC reboots → automatic Return-to-Home or Loiter.

• Stable Startup Handshake

  • Three-way acknowledgement chain (Drone ⇄ PX4, PX4 ⇄ MAVSDK, MAVSDK ⇄ GCS) to guarantee all SITL instances or hardware drones are ready before takeoff.
  • “OK-to-Start” broadcast from the cloud backend prevents premature launches.

• Robustness & Performance

  • Race-condition fixes under high CPU load (ensures correct GUIDED → AUTO transitions).
  • Emergency-land command reliability during mode transitions.
  • Network buffer tuning for large-scale simulations (100+ drones) using MAVLink Router.

• React Dashboard (Frontend)

  • Live position, battery, mode, and failsafe status for each drone.
  • Mission upload interface for offline trajectories or real-time swarm commands.
  • Formation editor (drag-and-drop) for offline shows (coming soon).
  • REST API endpoints via MAVLink2REST for external integrations.

• Automated Docker Environment

  • Prebuilt Docker image includes PX4 1.16, MAVSDK Drone Show, MAVLink Router, MAVLink2REST, Gazebo, and all SITL dependencies.
  • Auto hardware-ID detection and dynamic container creation scripts.

• Mission Configuration Tools

  • SkyBrush CSV → MDS template converter script.
  • JSON-based mission/formation files with built-in validators.
  • Parameter‐tuning utilities for leader election thresholds, Kalman filter gains, and failsafe timeouts.

---

Brief History & Changelog

• Version 0.1 (Mar 2020)

  • Single-drone CSV trajectory following.

• Version 0.2 (Jun 2020)

  • Multi-drone offset/delayed CSV trajectories.

• Version 0.3 (Oct 2020)

  • SkyBrush CSV processing; code optimizations for drone shows.

• Version 0.4 (Feb 2021)

  • Introduced Coordinator.py for advanced swarm; improved telemetry/commands.

• Version 0.5 (Jul 2021)

  • Basic leader/follower missions on hardware; enhanced GCS data handling.

• Version 0.6 (Dec 2021)

  • Complex leader/follower swarm control; Docker-based SITL environment.

• Version 0.7 (Apr 2022)

  • React GUI for real-time swarm monitoring; Docker automation for PX4 SITL.

• Version 0.8 (Sep 2022)

  • Major GUI enhancements; Kalman-filter–based swarm behaviors; optimized cloud SITL.

• Version 1.0 (Mar 2023)

  • Stable release; Flask web server; removed UDP dependencies.

• Version 1.5 (Aug 2023)

  • Mission config tools; SkyBrush converter; expanded MAVLink2REST.

• Version 2.0 (Nov 2024)

  • Enhanced React GUI, Flask backend, robust drone-show scripts, Docker SITL.
  • 100-Drone SITL Test Video

• Version 3.0 (Jun 2025)

  • Smart Swarm Leader–Follower fully operational (leader failover, auto re-election, seamless follower sync).
  • Global Mode Setpoints for unified offline & live missions.
  • Enhanced Failsafe Checks (preflight health checks, in-flight monitoring).
  • Stable Startup Sequence (three-way handshake, “OK-to-Start”).
  • Robustness & Bug Fixes (race conditions, emergency-land reliability, network buffer tuning).
  • Unified All-in-One System for both drone shows and live swarm.
  • Smart Swarm Clustered Leader–Follower Video
    

---

Getting Started

All the detailed setup instructions—downloading the Docker image, creating SITL instances, running both Drone-Show and Smart Swarm modes—live available in docs/sitl_demo_docker.md:

Note: That document covers:

• Docker image pull/load commands
• setup_environment.sh and create_dockers.sh usage
• Network, MAVLink Router, Netbird VPN setup
• React dashboard startup (linux_dashboard_start.sh --sitl)
• How to upload offline trajectories or launch live swarm missions

---

Contact & Contributions

If you run into issues, have suggestions, want to contribute new features or need help with your custom business requirements:

• Email: p30planets@gmail.com
• LinkedIn: Alireza Ghaderi

We welcome code, documentation improvements, Docker recipes, CVs, and new swarm algorithms. Please fork the repo, branch from v3.0, and submit pull requests with detailed descriptions.

---

Disclaimer

Using offboard mode or live swarm control on real drones carries risk. Ensure you have the necessary expertise, understand safety implications, and implement robust failsafe procedures before attempting any real-world flights. Always prioritize regulatory compliance and flight safety.

---

Additional Resources

• GitHub Repository (v3.0): https://github.com/alireza787b/mavsdk_drone_show

• SITL Demo with Docker Guide:

  • docs/sitl_demo_docker.md

• YouTube Tutorials:

  • Project History & Tutorials
  • IoT-Based Telemetry & Video Drone Concepts

• MAVSDK Documentation: https://mavsdk.mavlink.io/

• PX4 Autopilot: https://px4.io/

• Netbird VPN Docs: https://docs.netbird.io/

• MAVLink2REST Documentation: https://github.com/mavlink/mavlink2rest

• SkyBrush Drone Show Tool: https://skybrush.io/

---

© 2025 Alireza Ghaderi Licensed under CC BY-SA 4.0. Feel free to reuse or modify under the terms; please attribute and link back to the original repository.