Diver relative uuv navigation for joint human-robot operations

Abstract

A novel application for Autonomous Underwater Vehicles (AUVs) is considered here: a robotic diver assistant that enables close-quarters robotic operations with human divers. A robotic diver assistant has the potential to improve the efficiency, effectiveness and safety of diver operations. The robot diver assistant must share the operating environment with human divers, navigate relative to the environment to reach a specified site location (along with moving divers), and then maneuver among the mostly static divers as they perform their tasks on location. The robot operates in three unique scenarios: station-keeping, diver-following (shadowing), and diver-leading (vectoring). Various strategies for navigating among divers while ensuring diver safety are investigated. A reactive strategy, based on potential fields, is investigated and applied to station-keeping and diver-following. A deliberative approach, which plans the robots motion over a finite horizon, is presented for diver leading. These approaches are applied to the SeaBotix vLBV300 platform for which a simulator is developed based on a decoupled motion model for the platform, as well as experimental results in a controlled test tank.