摘要
分析了传统时间序列分析法建立ARMA模型的不足,提出了一种利用模型阶数判断准则和长自回归法建模的新方法。对数据进行了预处理和模型识别,比较了不同模型阶数判断准则的特点,研究了用长自回归法估计模型参数的步骤和方法。最后对光纤陀螺随机漂移进行了建模实验,通过对模型残差的白噪声检验,证实了使用新方法建立的ARMA模型具有很好的适用性。
The disadvantage of establishing ARMA model with traditional time series analysis is analyzed; a new model building method based on judgment rules and long autoregression is put forward. The data pretreatment is made, the model is identified, several and judgment rules of model order are compared and the steps and methods of using long autoregression to estimate the coefficients are studied. In the end, experimental modeling of the FOG random drift is carried out, through white noise checking of the model r...
出处
《装甲兵工程学院学报》
2008年第3期50-53,共4页
Journal of Academy of Armored Force Engineering
基金
军队科研计划项目
