Gazebo

Introduction 

For simulation purposes with Gazebo simulator , Aerostack2 provides with a platform that serves as an entry point for aerial robotics simulated in this environment.

Installation 

Prerequisites 

Gazebo in Fortress version is required. For installation instructions, follow the Gazebo Fortress installation guide.

Install platform package 

For binary installation, install by running:

sudo apt install ros-humble-as2-platform-gazebo

Warning

This package is not available for binary installation yet. Cooming soon.

For source installation, clone Aerostack2 repository into your workspace and build it. See it in Aerostack2 installation guide.

Install simulation assets 

For binary installation, install by running:

sudo apt install ros-humble-as2-gazebo-assets

Warning

This package is not available for binary installation yet. Cooming soon.

For source installation, clone Aerostack2 repository into your workspace and build it. See it in Aerostack2 installation guide.

Aerostack2 Common Interface 

For more details about platform control modes and sensors, see Aerostack2 Aerial Platform Concepts. For more details, about bridge between Gazebo and ROS, see the Aerostack2 Gazebo ROS Bridge.

Control Modes 

These are supported control modes:

Control Modes Gazebo Platform
Control Mode	Yaw Mode	Reference Frame
Hover	None	None
Speed	Speed	FLU

Sensors 

These are supported sensors:

Sensors Gazebo Platform
Sensor	Topic	Type
IMU	sensor_measurements/imu	sensor_msgs/msg/Imu
Magnetometer	sensor_measurements/magnetic_field	sensor_msgs/msg/MagneticField
Air pressure	sensor_measurements/air_pressure	sensor_msgs/msg/FluidPressure
Battery	sensor_measurements/battery	sensor_msgs/msg/BatteryState
Camera	sensor_measurements/{model_name}/image_raw	sensor_msgs/msg/Image
Depth camera	sensor_measurements/{model_name}/depth	sensor_msgs/msg/Image
Camera	sensor_measurements/{model_name}/camera_info	sensor_msgs/msg/CameraInfo
Lidar	sensor_measurements/{model_name}/scan	sensor_msgs/msg/LaserScan
Lidar	sensor_measurements/{model_name}/points	sensor_msgs/msg/PointCloud2
Camera	sensor_measurements/{model_name}/points	sensor_msgs/msg/PointCloud2

Gimbal 

Gimbal is supported in simulation. These are the supported gimbal model types:

Gimbal Control Modes Ignition Gazebo Platform
Gimbal type	Topic	Type	Control mode id
gimbal_position	platform/{gimbal_name}/gimbal_command	as2_msgs/msg/GimbalControl	“0”
gimbal_speed	platform/{gimbal_name}/gimbal_command	as2_msgs/msg/GimbalControl	“1”

Gimbal state is published in the following topics:

Gimbal State Ignition Gazebo Platform
Topic	Type
sensor_measurements/{gimbal_name}/twist	geometry_msgs/msg/Vector3Stamped
sensor_measurements/{gimbal_name}/attitude	geometry_msgs/msg/QuaternionStamped

Config Simulation 

There are two aerial models available for simulation. These models are:

Quadrotor base.

../../_images/quadrotor.png — Quadrotor base model

Hexrotor base.

../../_images/hexrotor.png — Hexrotor base model

In order to add an aerial model with sensors attached to it to the simulated world, Aerostack2 uses a configuration file, with JSON format, with the following structure:

{
  "world_name": "empty",
  "drones": [
      {
        "model_type": "quadrotor_base",
        "model_name": "drone_sim_0",
        "xyz": [ 1.0, 1.0, 0.3 ],
        "rpy": [ 0.0, 0.0, 0.0 ],
        "payload": [
            {
              "model_type": "gps",
              "model_name": "gps0"
            },
            {
              "model_name": "gimbal",
              "model_type": "gimbal_speed",
              "payload": {
                  "model_name": "hd_camera0",
                  "model_type": "hd_camera"
              }
            }
        ]
      }
  ]
}

Where:

world_name: name of the defined world in sdf format.
drones: list of drones to be included in the world.

Each of the drones is defined by:

model_type: model of the drone defined in sdf format.
model_name: namespace
xyz: spawn position
rpy: spawn orientation
payload: list of sensors/gimbal attached to the model

Each element of the payload is defined by:

model_type: name of the sensor/gimbal inside the simulation (this case gps and gimbal_speed)
model_name: name of the sensor/gimbal defined in sdf format.

If a drone payload contains a gimbal, a gimbal should contain a payload which must containt a sensor.

New models, sensors and worlds are defined in the as2_gazebo_assets package. For more information on how to create new assets, go to the Gazebo Fortress tutorial page.

Platform Launch 

Aerostack2 Gazebo platform provides a launch file, which parameters are:

Gazebo Platform Parameters
Parameter	Type	Description
namespace	string	Namespace of the platform, also named as drone id.
config_file	string	Path to the simulation configuration file.
control_modes_file	string	Optional. File path with the control modes configuration. Default the one in the package.
use_sim_time	bool	Optional. Syncronize simulation time with node time. Default false.
enable_takeoff_platform	bool	Optional. Enable takeoff by platfom. Default false.
enable_land_platform	bool	Optional. Enable land by platfom. Default false.

Example of launch command:

ros2 launch as2_platform_gazebo platform_gazebo_launch.py namespace:=drone_sim_0 config_file:=world_json_path

For launch the simulation, run the following command:

ros2 launch as2_gazebo_assets launch_simulation.py config_file:=world_json_path