A ROS Implementation of the Situational Awareness and Maneuvering Systems for an Autonomous Marine Vessel

摘要

This paper describes a study on the development, implementation, and evaluation of situational awareness, and the maneuvering systems for an autonomous ship. The implementation and validation of these modules within the Robot Operating System (ROS) framework are motivated by the participation of the Atlantics team (UPORTO) in the Njord - The Autonomous Ship Challenge, a pilot student-driven competition that involves the participation of several universities and is organized by a team with the Norwegian University of Science and Technology. The competition consists of a set of several tasks that the participating teams are asked to solve. These tasks are indeed current challenges in the marine industry towards autonomy and involve the design and evaluation of motion control and navigation algorithms with situational awareness and collision avoidance capabilities for a standardized vessel and sensor suite. In this paper, we present the modules of object detection, object tracking, and object state estimation that constitute the situational awareness system, and the modules for planning and motion control, including a spline-based motion planner, a nonlinear path-following controller with a safety critical module for obstacle avoidance, and a model reference adaptive controller for the inner-loop (autopilot). Several simulation results using the competition simulator are shown.