Drone-Navigation-Simulation-ROS2

Docker configuration for autonomous navigation in GPS-denied environment with PX4 Autopilot.

Details

• Auto-deletion: After its execution, the Docker container is automatically deleted. This helps in maintaining a clean environment by freeing up resources on the host machine.

• ros2_ws-src folder: This folder should contain the source code for the ROS2 workspace. The Docker container comes pre-loaded with px4-ros-com and px4-msgs, which are dependencies for the PX4-ROS2 communication. In this workspace there is the leonardo_field_description package which contains the world simulation.

• PX4 Firmware: The PX4 Firmware is not included in the Docker container. Instead, its folder is shared with the container using a Docker volume. This allows the firmware to be updated without having to rebuild the Docker container.

Installing

A step by step series of examples that tell you how to get a development environment running:

1. Clone the repository to your local machine.
2. Clone the PX4 Firmware with git clone --single-branch -b release/1.14 git@github.com:vinsco30/PX4-Autopilot-PRISMA.git --recursive
3. Build the docker imagege with cd docker && docker build -t px4-ros2-gz -f px4_gz_sim_dockerfile.txt .
4. Run the container with ./run_cnt.sh.

Running

If you want to run the simulation in the Contest world, you have to start the simulation with:

ros2 launch leonardo_field_description launch_leonardo_race_field.py 

Then to spawn the x500 drone with depth camera you can run

PX4_SYS_AUTOSTART=4002 PX4_GZ_MODEL=x500_depth PX4_GZ_MODEL_POSE="-9.5 -3.5 0.5" ./PX4-Autopilot-PRISMA/build/px4_sitl_default/bin/px4 

Alternative Running (developing Navigation)

During the development of the aerial navigation stack several simulations need to be runned, to simplify this you can use tmux.

First, clone into ros2_ws/src/pkg the package for the bringup of the TF tree for PX4 simulation drone-bringup , with the command

git clone https://github.com/vinsco30/drone-bringup.git

Then you can compile and source the workspace and run the simulation with

./start.sh

which will start a tmux session with 4 screen opening gz, rviz2 and running the tf tree and the Micro XRCE-DDS agent.

To stop the session you should kill the gz simulator and then type ctrl+b :kill-server.

Headless simulation

If you need to tun simulation with depth camera, you can use headless simulation without GUI. You can start it running the following script

./start_depth_headless.sh

Bird Diverter Installation Simulation

In addition, it is possible to tun the simulation for the bird diverter installation application. The scripts provided will start a simulation in gz-sim with the custom Flamingo drone as an adaptation of the classic X500 frame equipped with a tool composed by a fixed stick mounted underneath the drone and a force sensor located at the tip of the stick. Run:

./start_flamingo.sh 

or

./start_flamingo_headless.sh

To start the simulation. A depth camera (OakD-Lite) too is spawned to detect possible high voltage cables.

TODO: implement mapping with depth camera