ardupilot_gz

This project contains ROS 2 packages for simulating models controlled by ArduPilot SITL with DDS support in Gazebo.

The project is adapted from the ros_gz_project_template project.

Included packages

• ardupilot_gz_description - Contains the SDFormat description of the simulated system.

• ardupilot_gz_gazebo - Contains Gazebo specific code such as system plugins.

• ardupilot_gz_application - Contains ROS 2 specific code and configuration.

• ardupilot_gz_bringup - Contains launch files and high level utilities.

Prerequisites

• Install ROS 2 Humble
• Install Gazebo Garden
• Follow the Installing Build Dependencies section of AP_DDS's README

Install

1. Create a workspace folder

mkdir -p ~/ros2_ws/src && cd ~/ros2_ws/src

2. Get the project source

cd ~/ros2_ws/src
wget https://raw.githubusercontent.com/ArduPilot/ardupilot_gz/main/ros2_gz.repos
vcs import --recursive < ros2_gz.repos

3. Set the Gazebo version to Garden:

export GZ_VERSION=garden

4. Update ROS dependencies

cd ~/ros2_ws
source /opt/ros/humble/setup.bash
sudo apt update
rosdep update
rosdep install --rosdistro $ROS_DISTRO --from-paths src -i -r -y

5. Build

cd ~/ros2_ws
colcon build --cmake-args -DBUILD_TESTING=ON

6. Test

source ./install/setup.bash
colcon test --packages-select ardupilot_sitl ardupilot_dds_tests ardupilot_gazebo ardupilot_gz_applications ardupilot_gz_description ardupilot_gz_gazebo ardupilot_gz_bringup
colcon test-result --all --verbose

Usage

1. Source the workspace

source ~/ros2_ws/install/setup.sh

2. Launch the simulation

ros2 launch ardupilot_gz_bringup iris_runway.launch.py

3. Launch a GCS (MAVPorxy)

mavproxy.py --console --map

4. Inspect topics

$ ros2 topic list
/ap/battery/battery0
/ap/clock
/ap/navsat/navsat0
/ap/tf_static
/clicked_point
/clock
/goal_pose
/initialpose
/iris/odometry
/joint_states
/parameter_events
/robot_description
/rosout
/tf
/tf_static

Advanced variations

Here's a list of all the additional use cases from this repository:

1. Iris Maze

This simulation has an iris copter equipped with a 360 degrees 2d lidar in a maze world.

ros2 launch ardupilot_gz_bringup iris_maze.launch.py

Notes

1. Additional dependencies

ros_gz has a dependency on gps_msgs included in

git clone https://github.com/swri-robotics/gps_umd.git -b ros2-devel

When building from source add COLCON_IGNORE to gpsd_client as this package is not required and will not build on macOS.

2. sdformat_urdf

2.1. Library extension

On macOS the robot_state_publisher node cannot load the sdformat_urdf_plugin plugin unless the extension is changed:

cd ./install/sdformat_urdf/lib
ln -s libsdformat_urdf_plugin.so libsdformat_urdf_plugin.dylib

2.3. Model URIs

The sdformat_urdf plugin requires the <uri> element to use the package:// prefix for a resource to be located by RViz. At present this requires the models to be edited.

All occurrences of

`model://{model_name}`

must be replaced with

`package://{package_name}/models/{model_name}`

2.4. SDFormat environment variables

The sdformat_urdf plugin uses the sdformat13 libraries to parse the model description which relies on the environment variable SDF_PATH to resolve model resources. This is usually set in gz-sim7, however when using the plugins standalone, for instance in the bring-up launch files, SDF_PATH must be set otherwise the plugin will not resolve the models and their dependencies.

source ~/ros2_ws/install/setup.sh
export SDF_PATH=$GZ_SIM_RESOURCE_PATH

This is assigned in the iris.launch.py file as SDF_PATH is not usually set by the ament environment hooks.