FlightGoggles : Pure simulation of single agent

This page describes simulation of a single agent dynamics and imagery. In this scenario, a user can use a keyboard or joystick to operate the agent or interface their autonomy code to the dynamics simulation by publishing rateThrust messages to the appropriate topic.

Use one of the following launch files as appropriate.

To run example environment with joystick/keyboard teleoperation for multirotors

roslaunch flightgoggles teleopExample.launch

To run core multirotor simulation framework without teleoperation

roslaunch flightgoggles core.launch

To run example environment with joystick/keyboard teleoperation for car

roslaunch flightgoggles teleopExampleCar.launch

To run core car simulation framework without teleoperation

roslaunch flightgoggles coreCar.launch

The FlightGoggles binary might take up to 30 seconds to load. In the development build, asset loading to the GPU has not yet been optimized.

If you'd like to keep the FlightGoggles render alive between tests to avoid waiting for the renderer to load, you can pass the argument use_external_renderer:=1 to the appropriate launch file and run the renderer separately as follows:

rosrun flightgoggles FlightGoggles.x86_64

Default Controls

For human teleoperation, a Logitech F310 controller or normal QWERTY keyboard can be used. Please refer to Manual Flight Control.