Modularis: Modular Underwater Robot for Rapid Development and Validation of Autonomous Systems
Baker Herrin, Victoria Close, Nathan Berner, Joshua Herbert, Ethan Reussow, Ryan James, Cale Woodward, Jared Mindlin, Sebastian Paez, Nilson Bretas, Jane Shin
TL;DR
Modularis addresses the high cost, slow bring-up, and limited modularity of autonomous underwater robots by delivering a low-cost, modular AUV built on the BlueROV2 platform. The design features mechanical, electrical, and software modularity, including a three-tube battery enclosure, a plug-and-play sensor main board, and a dual-ROS software stack (ROS1 tethered and ROS2 untethered) with Gazebo-based simulation. Key contributions include the Main Sub-System (MMB), Power Sub-System with a buck converter and BMS, Sensor Sub-System with a CAN-enabled IMU, and a software framework that supports rapid testing of perception, planning, and control. The open-source hardware and software, combined with both tethered and untethered operation and Gazebo simulation, enables faster validation of autonomous algorithms and broader applicability to environmental monitoring and underwater mapping.
Abstract
Autonomous underwater robots typically require higher cost and time for demonstrations compared to other domains due to the complexity of the environment. Due to the limited capacity and payload flexibility, it is challenging to find off-the-shelf underwater robots that are affordable, customizable, and subject to environmental variability. Custom-built underwater robots may be necessary for specialized applications or missions, but the process can be more costly and time-consuming than purchasing an off-the-shelf autonomous underwater vehicle (AUV). To address these challenges, we propose a modular underwater robot, Modularis, that can serve as an open-source testbed system. Our proposed system expedites the testing of perception, planning, and control algorithms.
