Hardware-in-the-Loop Trajectory Tracking and Collision Avoidance of Automated Inland Vessels Using Model Predictive Control*

The automation of ships offers great potential in terms of safety and efficiency. For this reason, the research project AKOON investigates the automation of an overactuated river ferry. This task requires the development of advanced control algorithms that can safely track a roughly planned trajectory in spite of obstacles nearby, external disturbances, and limited propulsion forces. This paper presents a new control concept specifically designed for the automation of the overactuated river ferry. Since usual control approaches such as feedback linearization or linear model predictive control are not suited for the consideration of most real obstacles, a nonlinear model predictive control using a multiple shooting strategy is proposed. In addition, a disturbance observer is presented that utilizes a customized dynamic model of the examined ferry. The control concept is implemented on an industrial controller that is examined in a hardware-in-the-loop test bench resembling a real-world configuration. During the river crossing, a maximum horizontal tracking error of less than 5 m and a maximum yaw angle error of less than 10° are desired whenever the reference trajectory is feasible. It is shown by means of an exemplary simulation scenario that the proposed control concept enables a collision-free trajectory tracking and that the specified maximum errors are not exceeded.