| 引用本文: | 茅宁 ,张良起.多维滑动和（MA）模型参数的自适应估计[J].控制理论与应用,1990,7(4):29~36.[点击复制] |
| Mao Ning, Chang Liangqi.The Adaptive Estimation of the Parameters of a Vector Moving Average (MA) Model[J].Control Theory and Technology,1990,7(4):29~36.[点击复制] |
|
| 多维滑动和（MA）模型参数的自适应估计 |
| The Adaptive Estimation of the Parameters of a Vector Moving Average (MA) Model |
| 摘要点击 921 全文点击 390 投稿时间：1988-12-28 修订日期：1989-07-08 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 |
| 1990,7(4):29-36 |
| 中文关键词 系统辨识 参数估计 LS梯格滤波 滑动和模型 递推算法 |
| 英文关键词 system identification parameter estimation moving average (MA) model LS ladder filtering recursive algorithm |
| 基金项目 |
|
| 中文摘要 |
| 本文利用AR模型LS梯格滤波的有关公式给出了实时辨识多维MA模型参数的递推算法。该算法建立在对模型噪声和观测的自协方差阵和互协方差阵的矩估计基础上，由一个N阶反馈形式的梯格滤波器构成，可关于时间和阶次双重递推。该算法计算量为O(N)的量级，并具备梯格滤波固有的良好数值及结构特性。 |
| 英文摘要 |
| A new recursive algorithm for the on line identification of the parameters of a vector moving average (MA) model is presented by using the results of the LS ladder filtering for an autoregressive (AR) model. It is based on the moment estimation of auto-and cross-covariance matrice of the model noise and observation. The scheme can be represented in a feedback ladder form of order N, recursive both in time and order. The algorithm requires O(N) operations per input sample and preserves the advantages of the nice numerical and structural properties of the LS ladder algorithm. |