Hexacopter Trajectory Tracking using MPC and LOS Guidance

This repository contains the implementation of a real-time trajectory tracking framework for a hexacopter UAV using Model Predictive Control (MPC) combined with Line-of-Sight (LOS) guidance.
The project was developed as part of an undergraduate thesis and focuses on bridging advanced control theory with practical real-world implementation on a PX4-based UAV platform.

Overview

The proposed control architecture follows a hierarchical control structure:

• LOS guidance is used to generate smooth and continuous reference trajectories.
• MPC is implemented as an outer-loop controller to regulate the translational motion of the UAV.
• Attitude and angular rate stabilization are handled by the built-in PID controllers of the PX4 autopilot.

The MPC controller runs offboard on a companion computer and communicates with PX4 via MAVROS.
The system is validated through both Software-In-The-Loop (SITL) simulation and real-time flight experiments on a physical hexacopter.

System Architecture

• Path reference → LOS guidance
• LOS output (position, velocity, yaw reference) → MPC outer loop
• MPC output (translational acceleration) → attitude and thrust setpoints
• PX4 inner loop (PID) → motor commands

This structure allows the use of advanced predictive control while maintaining compatibility with a commercial autopilot.

Tools and Technologies

• PX4 Autopilot
• ROS (Robot Operating System)
• MAVROS
• Gazebo (SITL simulation)
• Raspberry Pi (companion computer)
• Model Predictive Control (MPC)
• Line-of-Sight (LOS) guidance

Key Features

• Real-time MPC-based outer-loop control for hexacopter UAV
• LOS guidance for smooth 3D trajectory tracking
• Helical trajectory tracking in three-dimensional space
• Simulation-to-hardware workflow (SITL → real flight)
• Comparative evaluation against PX4 PID position control

Experimental Results

Real-time flight experiments demonstrate that the proposed MPC–LOS framework:

• Achieves stable and smooth tracking of a 3D helical trajectory
• Reduces trajectory tracking RMSE by 9.46%
• Improves mission completion time by 18.2% compared to the conventional PX4 PID-based position controller.

Notes

• This repository focuses on trajectory tracking, not low-level motor control.
• MPC parameters are tuned for real-time feasibility on embedded hardware.
• The code is provided for research and educational purposes.