ROS2 Dependable MPC for UAV Control

This repository provides the implementation of a Model Predictive Control (MPC) architecture designed for the reliable control of unmanned aerial vehicles (UAVs). It is part of the Master's thesis "Design and Implementation of a Model Predictive Control Architecture for Dependable Control of Unmanned Aerial Vehicles" by Frederik Schmittel.

Key Technical Features

Advanced Control System Implementation

• Model Predictive Control (MPC)

  • Custom ADMM-based solver for optimal control computation
  • Real-time constraint handling to ensure system safety

• State Estimation

  • Luenberger Observer for Robust State Estimation
    • Estimates system states without noise influence
    • Real-time state prediction and correction

Safety-Critical Systems Design

• Predictive Safety Filter
  • Proactively enforces constraints to ensure safe operation
  • Real-time validation of control inputs
  • Automatic fallback to a backup control input in case of constraint violations

Modern Software Architecture

• ROS2 Integration

  • Modular, node-based system architecture
  • Efficient inter-process communication through ROS2 topics
  • Zero-copy data transfer where applicable to minimize latency

• High-Performance Computing

  • Optimized matrix operations for real-time execution
  • Cache-friendly data structures to enhance computational efficiency
  • Minimization of dynamic memory allocation for improved determinism

The system consists of three main components working together to ensure safe and optimal control of a dynamic system.

graph TD
    A[System Simulator] -->|Measurements| B[State Observer]
    B -->|State Estimate| C[MPC Controller]
    C -->|Control Commands| D[Safety Filter]
    D -->|Safe Controls| A

Loading

Core Components

1. System Simulator (system_simulator.cpp)

• Simulates linearized UAV dynamics
• Provides sensor measurements
• Applies control inputs
• Real-time system state updates

2. Safe Controller (safe_control.cpp)

• Implements predictive safety filtering
• Validates control constraints
• Ensures state bounds
• Provides backup control strategies

3. Controller (controller.cpp)

• Executes MPC optimization
• Performs state estimation
• Computes optimal control inputs
• Handles real-time constraints

Performance Characteristics

• Control loop frequency: 50Hz
• State dimension: 8 (position, velocity, orientation)
• Control inputs: 4 (motor commands)
• Constraint validation: <1ms
• MPC solve time: ~5ms average

Getting Started

Prerequisites

• Docker
• Git
• ROS2 Humble (if building from source)

Quick Start with Docker

# Clone repository
git clone https://github.com/F-DSK-SCH/ros2_dependable_mpc.git
cd ros2_dependable_mpc

# Build Docker image
docker build -t ros2_dependable_mpc .

# Run simulation
docker run --rm ros2_dependable_mpc ros2 launch mpc_package control_loop_launch.py

Building from Source

# Install dependencies
sudo apt-get update && sudo apt-get install -y \\
    python3-pip \\
    python3-colcon-common-extensions \\
    libeigen3-dev

# Build project
mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
git clone https://github.com/F-DSK-SCH/ros2_dependable_mpc.git
cd ..
colcon build
source install/setup.bash

# Run simulation
ros2 launch mpc_package control_loop_launch.py

System Configuration

The system can be configured through YAML files:

controller:
  ros__parameters:
    # System dynamics matrix (8x8)
    SYSTEM_MATRIX: [
      1.0, 0.0, -0.0001, ...
      # ... rest of matrix
    ]
    
    # Input matrix (8x4)
    INPUT_MATRIX: [
      0.0, -0.0065, ...
      # ... rest of matrix
    ]
    
    # Control constraints
    CONTROL_MAX: [0.43633, 0.34907, 0.52360, 0.75]
    CONTROL_MIN: [-0.43633, -0.34907, -0.52360, 0.0]

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

This work is part of the Master's thesis "Design and implementation of a model predictive control architecture for dependable control of unmanned aerial vehicles" by Frederik Schmittel.

Project Link: https://github.com/F-DSK-SCH/dependable_mpc