Roll Stabilization by Rudder Control through Multi-variate Auto-Regressive Model

Abstract

In the conventional roll stabilization system, fins or anti-rolling tank has been utilized as the actuator. However, it does not furnish the function to keep the ship's course. On the otherhand, some researchers have concentrated on the study of rudder roll combined motion since 1970's and the development of the autopilot system for rudder roll stabilizatin since 1980's. The first part of this paper will be devoted to examine the feasibility of the rudder roll stabilization by statistical analysis of the data which were gained by full scale experiment. Based on these results, the authors propose a Rudder Roll Control System (RRCS) to control not only yawing motion but also rolling one simultaneously, using the modern optimal control theory. In this system, a multi-variate auto regressive model of the rudder, roll and yaw motion are fitted to the actual data which were obtained by a full scale ship's experiment and the optimal control gain under appropriate evaluating function are calculated. In the last part of this paper, the authors will report the results of the simulation to assess the effect of control and the full scale experiments by a small training ship.