船摇数据实时滤波与预报的时序法

Time-Sequence Method for the Real-Time Filtering and Prediction of the Ship-Swaying Data

作者为船摇数据建立一个时序时域模型——长自回归模型 ,在此模型基础上给出了船摇数据的实时滤波和预报的方法 ,并从均方误差、预报误差、残差序列的相关性、拟合方差、仿真计算等多方面考察了滤波及预报效果。AR(p)模型的系数估计采用最小二乘递推方法 ,用较少的运算量和存贮量 ,得到了较高的估值精度。

This thesis builds a model for timesequence and timefield(a model of long autoregression).Using this model,a method of realtime filtering and prediction for the shipswaying data is acquired.We also reviewed the effect of the filtering and prediction from the mean square errors,predicated errors,the dependency of the remained errors sequence,conformed square errors,emulated calculations and so on.To get the coefficient estimation of the AR(p) model,we used the Least Square method and calculated by steps.We used a little calculation quantity and saving capacity to get a higher estimated precision than before.In fact,It is a very good method worthy of recommending to deal with the realtime data of the swaying ships.