Flight Demo

warning

Make sure you have completed the Hardware Setup Guide before running this demo.

This demo showcases the Isaac ROS VSLAM node in action, allowing a quadcopter to autonomously fly a specified pattern using visual SLAM for pose estimation. The quadcopter will utilize the MAVROSPY node to receive pose estimates and control the flight pattern.

Procedure

To run the flight demo, start by launching the Isaac ROS docker container:

$ cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
  ./scripts/run_dev.sh -b

Inside the container, navigate the to the vslam directory:

$ cd ${ISAAC_ROS_WS}/VSLAM-UAV/vslam

Next, launch the Issac ROS VSLAM node:

$ ./vslam_launch.sh

In a new terminal, launch the mavrospy docker container:

$ cd ${ISAAC_ROS_WS}/VSLAM-UAV/docker/mavrospy && \
  ./run_docker.sh

In the mavrospy container, launch the mavrospy node:

$ cd ~/VSLAM-UAV/vslam
$ ros2 launch mavrospy.launch.py

note

If you wish to ensure that the VSLAM pose estimation is being published correctly, you can run the following commands in a new terminal:

$ cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
  ./scripts/run_dev.sh
$ export ROS_DOMAIN_ID=1
$ ros2 topic echo /mavros/vision_pose/pose_cov

tip

You can also check that the flight controller is properly interpreting the pose estimates by test flying in POSITION flight mode.

Finally, switch the quadcopter to OFFBOARD mode and watch it fly!

info

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

$ ros2 launch mavrospy.launch.py pattern:=figure8

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.

Visualization

If you wish to visualize the VSLAM output in RViz, you can run the following commands in a new terminal:

$ cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
  ./scripts/run_dev.sh -b
$ export ROS_DOMAIN_ID=1
$ rviz2 -d ${ISAAC_ROS_WS}/VSLAM-UAV/vslam/vslam_realsense.cfg.rviz

warning

Running RViz on a remote desktop is generally not recommended due to performance issues. Even working with X11 forwarding can be slow depending on your network connection.

Instead, consider recording the demo using the ros2 bag record -a command and then visualizing the recorded data later.

Demo Video

note

The motion capture cameras visible in the video are used solely for ground truth pose validation. The UAV relies entirely on visual-inertial odometry (VIO) and the flight controller's IMU, fused through an Extended Kalman Filter (EKF), for local pose estimation during flight.