Direct Estimation Method for the Ship Motion Parameters based on Time Series Analysis

Daisuke Terada, Genshiro Kitagawa
2009 Journal of the Japan Society of Naval Architects and Ocean Engineers  
In this paper, a statistical estimation procedure for parameters in an equation of ship's oscillation, such as a damping coefficient and a natural frequency, is proposed. It is important that a mariner grasps the performance of the ship, when carrying out safe operation of the ship. From a viewpoint of avoiding a resonance with a wave, it is especially important for the mariner to grasp the natural frequency of the ship. Using the procedure, the parameters can be estimated by using an observed
more » ... using an observed data on board. Therefore, the mariner can always grasp the parameters of the ship under navigation. In this procedure, two kinds of stochastic techniques are adopted. The first one is a CAR (Continuous AutoRegressive) model that can gain precise parameters of ship's oscillating equation. In general, the ship motions in waves are regarded as a stochastic process. However, it has the problem that the noise process can not regard as a white noise. In the CAR model, it can be coped with by considering that the noise is an auto regressive process. The second one is a self-organizing state space modeling that can evaluate both the state estimation and the identification of unknown parameters in a general nonlinear non-Gaussian system. In this estimation calculation, Monte Carlo Filter which is a kind of Particle Filter is adopted. The goodness of the model is evaluated by using AIC (Akaike Information Criterion) with respect to the model. To examine the reliability of the proposed procedure, the numerical experiments are carried out by using the reproduced data under the condition which the unknown parameters are a fixed value. The estimated ship motion parameters based on the proposed method shows good agreement with the input data. Moreover, to examine the proposed procedure, the onboard experiments are carried out. Although some future problems became clear, we conclude that the proposed method is a powerful tool for the estimation of the ship motion parameters.
doi:10.2534/jjasnaoe.9.127 fatcat:pb7baqdw6nc3nma4kwzxmcriga