RoboWhaler: A Robotic Vessel for Marine Autonomy and Dataset Collection

摘要

In this paper, we present a manned/unmanned vehicle that is used for autonomous surface vessel algorithmic development. This mobile marine autonomy laboratory is built upon a modified Boston Whaler and can be operated in both manual and fully autonomous modes. The primary driver in development was the creation of a marine autonomy sensor suite capable of high resolution sensing in all weather conditions. The vessel now serves as a mobile marine autonomy laboratory and is used to collect on-water datasets as well as testing behaviors and algorithms for autonomous operations. The vessel is equipped with a SBG Ellipse2-D RTK INS, a Velodyne VLP-16 lidar for high resolution 360° sensor coverage, a Simrad 4G broadband radar for long-range 360° sensor coverage, forward facing FLIR ADK infrared cameras, and forward facing FLIR Blackfly 1.3 MP cameras. All data is available in ROSBag format and can be manipulated and visualized using the open source Robot Operating System (ROS). We also make available tools to extract timestamped data presented in standard human readable formats. Datasets have been collected with the current marine autonomy sensor suite from fall 2019 to summer of 2021. These datasets have been curated for consumption and are available on the MIT Sea Grant website https://seagrant.mit.edu/auvlabdatasets-marine-perception-1.