PX4 MPC - Interfacing PX4 with Model Predictive Control

This package contains an MPC integrated with with PX4 Autopilot and ROS 2.

The MPC uses the acados framework

Citing PX4-MPC

If you find this package useful in an academic context, please consider citing the paper

• Roque, Pedro, Sujet Phodapol, Elias Krantz, Jaeyoung Lim, Joris Verhagen, Frank Jiang, David Dorner, Roland Siegwart, Ivan Stenius, Gunnar Tibert, Huina Mao, Jana Tumova, Christer Fuglesang, Dimos V. Dimarogonas. “Towards Open-Source and Modular Space Systems with ATMOS.” arXiv preprint arXiv:2501.16973 (2025). . [preprint]

@article{roque2025towards,
  title={Towards Open-Source and Modular Space Systems with ATMOS},
  author={Roque, Pedro and Phodapol, Sujet and Krantz, Elias and Lim, Jaeyoung and Verhagen, Joris and Jiang, Frank and Dorner, David and Siegwart, Roland and Stenius, Ivan and Tibert, Gunnar and others},
  journal={arXiv preprint arXiv:2501.16973},
  year={2025}
}

Setup

The MPC formulation uses acados. In order to install acados, follow the following instructions. After building acados remember to install the python interface as described here.

To build the code, do the following steps:

mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
git clone https://github.com/PX4/px4_msgs.git  # PX4 messages for communication with PX4 DDS
git clone https://github.com/Jaeyoung-Lim/px4-offboard.git  # Rviz interface and setpoint generation
git clone https://github.com/DISCOWER/px4-mpc.git  # this package
cd ..
colcon build --packages-up-to px4_mpc
source install/setup.bash  # use source install/setup.zsh for zsh users 

Running MPC with PX4 SITL

In order to run the SITL(Software-In-The-Loop) simulation, the PX4 simulation environment and ROS2 needs to be setup. For instructions, follow the documentation

Before proceding, start the DDS interface of PX4 by running

micro-xrce-dds-agent udp4 --port 8888

Quadrotor Example

On a new terminal, navigate to your PX4-Autopilot directory and run the following command to start the PX4 SITL with Gazebo:

cd ~/PX4-Autopilot
make px4_sitl gz_x500

On another terminal, start PX4-MPC

ros2 launch px4_mpc mpc_quadrotor_launch.py 

Now, check the QGC or Headless setup in section QGC Setup or Headless (no-QGC) Setup at the bottom of this guide.

Then, in the same terminal where you started the PX4 SITL, you can control the quadrotor using the following commands:

• To arm the vehicle:

commander arm

• To follow the MPC setpoints:

commander mode offboard

Spacecraft Example

First, make sure that you have followed the instructions in the ATMOS guide. ATMOS MPC node also requires vehicle angular velocity data. Please modify the file PX4-Autopilot/src/modules/uxrce_dds_client/dds_topics-yaml and uncomment lines 56-57 to enable:

  - topic: /fmu/out/vehicle_angular_velocity
    type: px4_msgs::msg::VehicleAngularVelocity

Then, on your PX4-Autopilot directory, run the following command to start the PX4 SITL with Gazebo:

cd ~/PX4-Autopilot
make px4_sitl_spacecraft gz_atmos

On another terminal, run

ros2 launch px4_mpc mpc_spacecraft_launch.py mode:=wrench setpoint_from_rviz:=False

Then, in the same terminal where you started the PX4 SITL, you can control ATMOS using the following commands:

• To arm the vehicle:

commander arm

• To follow the MPC setpoints:

commander mode offboard

File truncated at 100 lines see the full file