Sonar based underwater vehicle dataset

This dataset are obtained from a BlueROV underwater vehicle in an indoor tank for the research of underwater perception, featuring ground-truth pose trajectories obtained from a motion capture system(MCS). In particular, we teleoperated the BlueROV and recorded readings from various sensors at $10$--$100$~Hz, including a high-resolution multibeam imaging sonar, an inertial measurement unit, and a Doppler velocity log (which also provides depth measurements). In addition, the underwater motion capture system provided the ground-truth trajectories of the BlueROV at $53$~Hz, which is the key enabler to effectively quantify the performance of underwater simultaneous localization and mapping (SLAM) algorithms.The dataset contains three folders and a configuration file. Each folder contains a Rosbag dataset, which are circle, line and L-shape trajectory data respectively. Configuration file is the transformation matrix between each sensor. Our dataset was collected on Ubuntu 20.04 with ROS version noetic 1.15.14.