Robotics

OpenMantaClaus: Open-source ROS2 AUV wins 2nd place on $1K budget

A Raspberry Pi-powered AUV with a USB webcam beats 20+ teams at SAUVC

Deep Dive

OpenMantaClaus is a complete, budget-friendly autonomous underwater vehicle (AUV) designed by independent developer Kushion for the Singapore Autonomous Underwater Vehicle Challenge (SAUVC). Despite costing a fraction of typical competition robots, it secured 2nd place against teams with far more expensive hardware. The entire autonomy stack runs on a Raspberry Pi with a Blue Robotics Navigator Flight Controller, and all vision processing happens through a single standard USB webcam.

Software-wise, the project is built on ROS 2 and extends the BlueOS ecosystem, using a forked BlueOS Extension to run containerized nodes for easy deployment and lifecycle management. Vision relies on custom YOLO-based object detection combined with classical computer vision techniques, all trained on a public Kaggle dataset the author released. Hardware uses low-cost Blue Robotics and ROVMaker components, and the full design is open source: code on GitHub, CAD on GrabCAD, and the YOLO dataset on Kaggle. It's an impressive proof that a small budget and clever engineering can compete at a high level.

Key Points
  • Runs full autonomy stack on Raspberry Pi + Blue Robotics Navigator, with only a USB webcam
  • Placed 2nd at SAUVC (Singapore AUV Challenge) on a tight budget
  • Open source: ROS 2 fork of BlueOS extension, YOLO dataset on Kaggle, CAD on GrabCAD

Why It Matters

OpenMantaClaus slashes AUV development costs, making underwater robotics research accessible to hobbyists and students worldwide.

📬 Get the top 10 AI stories daily