| 引用本文: | 衷路生, 宋执环.基于自适应密度估计的系统参数辨识[J].控制理论与应用,2007,24(5):851~855.[点击复制] |
| ZHONG Lu-sheng, SONG Zhi-huan.Parameter identification based on adaptive density estimation[J].Control Theory and Technology,2007,24(5):851~855.[点击复制] |
|
| 基于自适应密度估计的系统参数辨识 |
| Parameter identification based on adaptive density estimation |
| 摘要点击 1813 全文点击 942 投稿时间：2005-10-09 修订日期：2006-07-04 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/j.issn.1000-8152.2007.5.031 |
| 2007,24(5):851-855 |
| 中文关键词 自适应密度估计 非参数密度估计 参数估计 极大似然函数 |
| 英文关键词 adaptive density estimation non-parametric density estimation parameter estimation maximum likelihood function |
| 基金项目 国家973计划项目(2002CB312203-02). |
|
| 中文摘要 |
| 针对噪声分布未知的ARMAX系统, 提出了一种自适应非参数噪声密度估计方法, 由估计误差动态调整高斯核函数的全局带宽和局部带宽, 实现了未知噪声分布密度的自适应估计; 通过极小化似然函数, 给出了基于噪声密度估计的参数辨识迭代算法, 分析了算法的收敛性并给出了算法收敛的充分条件. 仿真结果表明本文提出的算法在系统噪声未知时具有较强的抗噪能力和良好的收敛性. |
| 英文摘要 |
| The problem of estimating parameters of ARMAX-model subject to unknown noise density distribution is addressed in this paper. Firstly, a method for adaptive nonparametric noise density estimation by using the Gaussian kernel is proposed. The global bandwidth and local bandwidth of the Gaussian kernel are dynamically computed according to the estimation error. Secondly, an iterative algorithm for parameter estimation is presented by minimizing the maximum likelihood function and the algorithm is based on the nonparametric estimate of the noise density. Thirdly, the convergence of the algorithm is analyzed and the convergent condition is given. Finally, simulation results show the proposed technique offers improved performance over existing methods when the noise density distribution is unknown. |