Skip to main content

Simulation

An additional benifit of using ROS is that it allows users to easily develop simulators such as Gazebo with their robotic projects. This is crucial for rapid development as it allows you to test your scripts without having to risk crashing your physical drone.

Docker​

Setting up Gazebo ROS on Linux is made easy via Docker.

This tutorial will assume you are running on a Linux OS, but Docker can be used on both Mac and Windows via Docker Desktop.

First, make sure you installed Docker Engine. It is also recommend you follow the Linux Post-Installation Steps as well to manage Docker as a non-root user.

We can verify our Docker installation via:

$ docker run hello-world

NVIDIA GPU Acceleration​

We can use Gazebo Classic to simulate our MAVROSPY movment commands. Using an NVIDIA graphics card will improve simulation performance, but is by no means required if you have a modern integrated graphics card.

To use our NVIDIA GPU, verify you have installed its driver:

$ nvidia-smi

If it's not installed, follow your desktops OS driver installation instructions.

To utilize our NVIDIA GPU inside a Docker container, we must install the NVIDIA Container Toolkit.

This is a summary of the official installation instructions.

Configure the production repository:

$ curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

Update the packages list from the repository:

$ sudo apt update

Install the NVIDIA Container Toolkit packages:

$ sudo apt install nvidia-container-toolkit

Configure the Docker container runtime by using the nvidia-ctk command:

$ sudo nvidia-ctk runtime configure --runtime=docker

Restart the Docker daemon:

$ sudo systemctl restart docker

Verify the installation by running a sample CUDA container:

$ sudo docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi

Depending on your PC, you may have an internal graphics card. To use your NVIDIA graphics card, enter this command:

$ sudo prime-select nvidia

Then, reboot.

Build Docker Image​

We can now build our Docker image via the following commands:

$ cd ~
$ git clone https://github.com/bandofpv/mavrospy.git
$ cd ~/mavrospy/docker/sim
$ bash run_docker.sh

This script will automatically build the Docker image and open it in a Docker container.

Launch Gazebo Classic Simulator​

Inside the docker container, launch a simulation:

$ roslaunch mavrospy gazebo_sim.launch
info

You can specify the flight pattern via the pattern argument. If unspecified, it will default to square.

$ roslaunch mavrospy gazebo_sim.launch pattern:=spiral

Available Patterns: square, square_head, circle, circle_head, figure8, figure8_head, spiral, and spiral_head. Where _head will specify the drone to face in the direction of motion.

Gazebo will open up and display a quadcopter model.

In the container, wait for the messages to stop updating until you see the following line:

INFO  [commander] Ready for takeoff!

Press Enter to open the PX4 shell. You should see a pxh> before your cursor.

Enter the following command to change into offboard mode:

pxh> commander mode offboard

Go back to the Gazebo window and watch drone fly!

The ROS program will automatically clean up after the drone lands. Press CTRL+D to exit/close the container.

Visualization with RViz​

We can also visualize our simulation poses with RViz.

After launching the Gazebo Classic Simulator, we can launch RViz.

Open a new termial and attach to the MAVROSPY sim container:

$ bash ~/mavrospy/docker/sim/run_docker.sh

Launch RViz:

$ roslaunch mavrospy rviz.launch

ROS will automatically begin plotting the drone's path.

CTRL+C to exit.