Towards adjustable autonomy for human-robot interaction in marine systems

Abstract

Current research into adjustable, or sliding, autonomy shows that human-in-the-loop robotic systems can greatly improve the efficacy of traditionally teleoperated or fully autonomous systems and the work presented here begins to define a framework of interfaces and controllers in support of multiagent human-robot teams in the marine environment. We present a scenario in which a harbor pilot, assisted by a team of USVs and a ground station operator, brings a ship into port and dock. To this end, we have developed a Robot Operating System (ROS) based Human-Machine Interface (HMI) focusing on the ground station operator and a harbor pilot tablet HMI concept to bridge the gap between current harbor pilot computer interfaces and the tools needed for the pilot to work in a human-robot team. Our preliminary simulations focus on a series of interactions within the sliding autonomy framework using a supervisory switching controller (SSC) as a task server with the human operator utilization ratio as a single feed-back signal. Our preliminary results show that the utilization ratio is sufficient to characterize the SSC for the human-in-the-loop system and appropriately adjust the level of autonomy.