船摇数据的时间序列模型处理

Time Series Model Processing of Ship Swaying Data

采用时间序列法可以弥补最小二乘法难以判别连续多个异常点的缺陷。首先,利用递推的方法得到模型的阶数和系数,然后分别使用最小二乘法、时间序列法处理船姿的纵摇数据,最后分析判断两种方法在探测出异常点的可行性和预测的精确性。

At present,LMS is commonly used,but it misjudges continuous abnormal data points.Meanwhile time series model can find out that outliers.In this paper,time series model is defined firstly,secondly the model order and coefficient is get by recursion,thirdly,the LMS method and time series model method is applied to ship swaying data.Finally the performance of both methods is shown,and some advice about ship swaying data processing is given according to simulation result.