Orca4 is a set of ROS2 packages that provide basic AUV (Autonomous Underwater Vehicle) functionality for the BlueRobotics BlueROV2.
Orca4 uses ArduSub as the flight controller and mavros as the GCS.
Orca4 runs in Gazebo Garden using the standard buoyancy, hydrodynamics and thruster plugins. The connection between ArduSub and Gazebo is provided by ardupilot_gazebo.
The BlueROV2 provides the following interesting sensors:
- An external barometer provides depth
- An IMU provides attitude
Orca4 adds a simulated down-facing stereo camera and ORB_SLAM2 to generate a 3D pose as long as the camera has a good view of the seafloor. The pose is sent to ArduSub and fused with the other sensor information.
If there is no view of the seafloor a synthetic pose is generated based on the last good pose and a simple motion model.
See orca_base for details.
Orca4 uses the Navigation2 framework for mission planning and navigation. Several simple Nav2 plugins are provided to work in a 3D environment:
- straight_line_planner_3d
- pure_pursuit_3d
- progress_checker_3d
- goal_checker_3d
See orca_nav2 for details.
See the Dockerfile for installation details.
Install these packages:
Build ArduSub for SITL:
cd ~/ardupilot
./waf configure --board sitl
./waf sub
Populate the workspace:
mkdir -p ~/colcon_ws/src
cd colcon_ws/src
git clone https://github.com/clydemcqueen/orca4
vcs import < orca4/workspace.repos
Build ros_gz for Garden, not Fortress:
export GZ_VERSION=garden
Get dependencies, ignoring Gazebo Garden rosdep keys:
rosdep update
rosdep install -y --from-paths . --ignore-src --skip-keys="gz-transport12 gz-sim7 gz-math7 gz-msgs9"
MAVROS depends on GeographicLib, and GeographicLib needs some datasets:
sudo mavros/mavros/scripts/install_geographiclib_datasets.sh
Build the workspace:
cd ~/colcon_ws
colcon build
In a terminal run:
source src/orca4/setup.bash
ros2 launch orca_bringup sim_launch.py
This will bring up all of the components, including the Gazebo UI. The surface of the water is at Z=0 and the sub will be sitting at the surface. The world contains a sandy seafloor 10 meters below the surface.
You should see ArduSub establish a connection to the ardupilot_gazebo plugin:
[ardusub-2] JSON received:
[ardusub-2] timestamp
[ardusub-2] imu: gyro
[ardusub-2] imu: accel_body
[ardusub-2] position
[ardusub-2] quaternion
[ardusub-2] velocity
At this point SLAM is not running because the seafloor is too far away, but the sub can still move using dead-reckoning. The base_controller node will send default camera poses to ArduSub to warm up the EKF and the manager node will request attitude information at 20Hz. Initialization completes when there is a good pose from the EKF:
[mavros_node-8] [INFO] [mavros.imu/handle_attitude_quaternion]: IMU: Attitude quaternion IMU detected!
[manager-9] [INFO] [manager/operator()]: EKF is running
[base_controller-10] [INFO] [base_controller/change_state]: EKF is running, state => RUN_NO_MAP
[base_controller-10] [INFO] [base_controller/UnderwaterMotion]: initialize odometry to {{-2.52304e-05, -3.28182e-05, -0.228547}, {0, 0, -5.00936e-05}}
Execute a mission in a second terminal:
source src/orca4/setup.bash
ros2 run orca_bringup mission_runner.py
The default mission will dive to -7m and move in a large rectangle. At -6m the cameras will pick up a view of the seafloor at and ORB_SLAM2 will start:
[orb_slam2_ros_stereo-13] New map created with 571 points
[base_controller-10] [INFO] [base_controller/change_state]: map created, state => RUN_LOCALIZED
You should notice a loop closure sometime during the 2nd run around the rectangle. The adjustment is very small.
[orb_slam2_ros_stereo-13] Loop detected!
[orb_slam2_ros_stereo-13] Local Mapping STOP
[orb_slam2_ros_stereo-13] Local Mapping RELEASE
[orb_slam2_ros_stereo-13] Starting Global Bundle Adjustment
[orb_slam2_ros_stereo-13] Global Bundle Adjustment finished
[orb_slam2_ros_stereo-13] Updating map ...
[orb_slam2_ros_stereo-13] Local Mapping STOP
[orb_slam2_ros_stereo-13] Local Mapping RELEASE
[orb_slam2_ros_stereo-13] Map updated!