2023_February_research_sprint_CAP

Abbreviations and nomenclature

	Explanation
ASV	Aquatic surface vehicle
ROV	Remote operated vehicle
DOF	Degree of freedom
CAP	Collebrative aquatic positioning
PnP	Perspective-n-Point
DS	Depth sensor
	Coordinate frame of A
	Translation from frame A to frame B
	Quaternion describes a rotation from frame A to frame B

Hardware setup

Robotic platform

Fig.1: BlueROV2	Fig.2: MallARD_003

In this experiment, BlueROV2 is the underwater robot tracked by CAP and Qualisys. MallARD_003 is as the ASV in CAP system, carried with a 2D LiDAR, monocular camera and an IMU and Ressperry Pi4 is the on-board PC.

Sensors

	Model number
IMU	Park Lord 3DMGQ7
LiDAR	TIM871 series Laser Scanner LiDAR
Presure sensor	Bar30 High-Resolution 300m depth/pressure sensor
USB camera	Low-Light HD USB camera

Software setup

IP address

Software on BlueROV

-......

Software on MallARD

MallARD_003 - SLAM on base

Datasets

There are four datasets compressed in Feb_2023_CAP1.zip. They are:

• 2023-02-24-16-11-29.bag
• 2023-02-24-16-21-23.bag
• 2023-02-24-16-31-55.bag
• 2023-02-24-16-57-03.bag

Datasets explenations

ROSBAG name	Description	Notes
2023-02-24-16-11-29.bag	MallARD moving BlueROV2 still	MallARD was controlled by joystick.
2023-02-24-16-31-55.bag	MallARD still BlueROV2 moving	BlueROV2 was on manual control and MallARD was hold still by an "apple picker"
2023-02-24-16-21-23.bag	Both moving 1	The movement of the two is not correlated.
2023-02-24-16-57-03.bag	Both moving 2	The movement of the two is not correlated.

Each dataset contains 8 rostopics:

ROSTOPIC name	Content
/bluerov/mavros/global_position/rel_alt	Depth sensor reading (parpendicular distance between water surface to depth sensor)
/bluerov/qualisys/blueROV/pose	Qualisys tracking BlueROV2's 6DOF (x, y, z, roll, yaw, pitch (in quaternion)) with respect to Qualisys world coordinate frame
/bluerov/usb_cam/image_raw	Image raw from camera mounted on BlueROV2
/imu/data	3DMGQ7 IMU reading (3-axis angular rate, 3-axis acceleration and rotation (in quaternion format) relative to boot-up frame with z axis point up)
/nav/filtered_imu/data	3DMGQ7 IMU reading (with build-in filter filtered)
/slam_out_pose	SLAM reading (MallARD's position (x, y) and orientation (yaw) relative to the map coordinate frame)
/tag_corners	Four conners' and centre's pixel coordinates in image plane
/tag_detections	ID of each detected tag, its pose (i.e., its position and orientation) in the camera frame.

Frames and offsets

Frames

Fig.3: Qualisys frame and map frame int the tank	Fig.4: Frames on MallARD	Fig.5: AprilTag fiducial marker and Qualisys object plate

Offsets

CAP requires some parameters to estimate the position of the marker, such as offsets between coordinate systems, the camera's intrinsic matrix, etc. In this experiment (system), multiple coordinate systems are involved. The static offsets between coordinate systems (including translations and rotations) are roughly measured and recorded in the table below. Note that these offsets may vary slightly when collecting each set of data.

Datasets name	Offsets	Notes
2023-02-24-16-11-29.bag	camera_intrinsic = [576.0913564252332, 0, 435.5688200267924; 0, 574.7447982699829, 285.3576802714121; 0, 0, 1] = [0, 0.7071, -0.7071, 0] = [0, 0, -0.19] = [0, 0, 0.7071, 0.7071] depth_sensor_offset = 0.2 = [-0.85, 0.75, 2.05] = [0.8660, 0, 0, -0.500] = [-0.1, 0, 0]	- B in superscript means Body frame. - C stands for camera frame. - I stands for IMU frame. - map frame is created by Hector SLAM. - depth_sensor_offset is the parpendicular distance from depth sensor to marker plane. - is the translation offset frome qualisys object center to AprilTag center (Fig.5)
2023-02-24-16-31-55.bag	camera_intrinsic = [576.0913564252332, 0, 435.5688200267924; 0, 574.7447982699829, 285.3576802714121; 0, 0, 1] = [0, 0.7071, -0.7071, 0] = [0, 0, -0.19] = [0, 0, 0.7071, 0.7071] depth_sensor_offset = 0.2 = [-0.8, 1.6, 2.05] = [0.8660, 0, 0, -0.500] = [-0.1, 0, 0]
2023-02-24-16-21-23.bag	camera_intrinsic = [576.0913564252332, 0, 435.5688200267924; 0, 574.7447982699829, 285.3576802714121; 0, 0, 1] = [0, 0.7071, -0.7071, 0] = [0, 0, -0.19] = [0, 0, 0.7071, 0.7071] depth_sensor_offset = 0.2 = [-0.7, 1.7, 2.05] = [0.8660, 0, 0, -0.500] = [-0.1, 0, 0]
2023-02-24-16-57-03.bag	camera_intrinsic = [576.0913564252332, 0, 435.5688200267924; 0, 574.7447982699829, 285.3576802714121; 0, 0, 1] = [0, 0.7071, -0.7071, 0] = [0, 0, -0.19] = [0, 0, 0.7071, 0.7071] depth_sensor_offset = 0.2 = [-0.96, 1.7, 2.05] = [0.8660, 0, 0, -0.500] = [-0.1, 0, 0]