Identification of a Maneuvering Vessel based on Regular Operation

Abstract

We present an approach for the identification of maneuvering vessels which is based on data collected during regular operations. In conjunction with previous work by the authors, this paper completes a comprehensive approach to estimating all parameters of a maneuvering model that fits exactly into the control system design concept developed as well. In this paper, the diagonal elements of the mass matrix are determined by evaluating acceleration periods of the vessel in the considered degrees of freedom. The contribution of the presented approach is the method of detecting and weighting all the available acceleration periods from a large database in order to automatically determine the parameters. Using a full-scale tugboat, the approach was tested and verified. In a real-time control loop application, dynamic positioning and a special tug maneuver were performed based on the estimation results from regular operation at fairly strong wind and current. The experiments indicate that the identification approach is extremely robust to disturbances from wind and current. This demonstrates the excellent performance of the entire system and validates the approach presented.