Project MUSK (Multi-UAV Simulation Kit)

A lightweight multi-UAV simulation kit built on PX4 + Gazebo + ROS Noetic, packaged in Docker. Provides a clean environment for testing algorithms without embedded implementations.

Features

• Simulation Framework Only: Pure simulation environment without algorithm implementations
• Multi-UAV Support: Configure multiple quadrotors with customizable formations
• Distributed Architecture: Each UAV runs as independent process with ROS-based communication
• State Machine Framework: Coordinated flight states (takeoff, guidance, landing) with synchronization
• Automated Workflow: One-click multi-UAV simulation launch with tmuxinator
• Dockerized: Fully containerized for reproducible environments
• Obstacle Scenarios: Built-in obstacle configuration for navigation testing

Tech Stack

• PX4: Open-source autopilot for UAV flight control
• Gazebo: 3D robotics simulation environment
• ROS Noetic: Middleware for communication and control
• MAVROS: Interface between ROS and MAVLink protocol
• Tmuxinator: Terminal multiplexer for automated multi-pane workflows
• Docker: Containerization for environment consistency

Quick Start

# Run simulation (using pre-built image)
./docker/run_docker.sh

# Launch multi-UAV simulation with tmuxinator
tmuxinator start . -p tmuxinator.yml scene=two_obstacles

# Exit: Ctrl+b then q in any pane

Build Custom Image

# Build your own image
./docker/build_docker.sh --push

Dockerfile provided for customization. Modify as needed for your requirements.

Distributed Framework

The project provides a sophisticated distributed multi-UAV control framework:

Architecture

• Each UAV runs as independent Python process
• ROS-based inter-UAV communication for state and telemetry
• Leader-follower coordination with configurable hierarchy

State Machine

• Coordinated flight phases: INIT → TAKEOFF → PREPARE → GUIDANCE → BACK → LAND
• Distributed synchronization between UAVs
• Automatic state broadcasting and monitoring

Safety Systems

• Real-time collision detection and avoidance
• Flight boundary enforcement (altitude, position limits)
• Emergency landing procedures

Flight Modes

• Position control (takeoff, landing, waypoints)
• Velocity control for formation flying
• Acceleration control for guidance algorithms
• Attitude control with throttle calibration
• Automatic hover throttle identification using binary search

Development Tools

• Tmuxinator for multi-pane development workflow
• Built-in data logging and rosbag recording
• 3D trajectory visualization tools
• Real-time obstacle visualization with RViz

Configuration

Simulation Scenes (two_obstacles.json)

• UAV count and formation parameters
• Leader-follower hierarchy
• Takeoff/prepare positions
• Obstacle locations and sizes

System Parameters (params.json)

• Safety boundaries and distances
• Throttle calibration settings
• Control gains and optimization parameters
• Execution timing configurations

Data Analysis & Visualization

# Plot trajectories and analyze flight data
python3 src/multi_uav_formation/multi_uav_formation/PlotSingleRun.py

Built-in tools for:

• 3D trajectory reconstruction
• Formation flying analysis
• Multi-UAV synchronization evaluation
• Performance metrics visualization

Roadmap

Planned Features

• ROS2 Support: Migrate from ROS Noetic to ROS2 for improved performance and modern DDS communication
• VTOL Support: Extend framework to support Vertical Take-Off and Landing (fixed-wing) aircraft

Requirements

• Docker only