| 祝徐轩,仲国民,何熊熊.多输入单输出时变输出误差模型学习辨识算法[J].高技术通讯(中文),2020,30(11):1140~1148 |
| 多输入单输出时变输出误差模型学习辨识算法 |
|
| |
| DOI：10.3772/j.issn.1002-0470.2020.11.006 |
| 中文关键词: 学习辨识； 最小二乘法； 随机梯度法； 辅助模型； 多输入单输出(MISO)系统 |
| 英文关键词: learning identification, least squares algorithm, gradient algorithm, auxiliary model, multiple-input single-output (MISO) system |
| 基金项目: |
|
| 摘要点击次数: 2416 |
| 全文下载次数: 1593 |
| 中文摘要: |
| 本文考虑了多输入单输出(MISO)时变输出误差系统参数的估计问题。对于多输入单输出时变输出误差系统，识别的难点在于待辨识的模型参量是随着时间而变化的，尤其突变的参数更难辨识。针对这一问题，本文将辅助模型的思想应用到学习算法中，给出了基于辅助模型的迭代学习随机梯度算法和基于辅助模型的迭代学习最小二乘算法的推导过程。最后，提供了说明性的仿真实例来分析所提出的算法，仿真结果表明基于辅助模型的迭代学习最小二乘算法可以快速跟踪突变的参数，获得精准的参数估计，验证了该算法的有效性。 |
| 英文摘要: |
| This paper presents a parameter estimation method for multiple-input and single-output (MISO) time-varying error model, whose parameters may change rapidly or abruptly with respect to time. For the system model undertaken, the problem in identification lies in that the model parameters to be identified change with time, and especially the rapid or abrupt change parameters are difficult to identify. With the aid of the auxiliary model, learning identification algorithms, including least-squares learning algorithm and the gradient learning algorithm, are proposed for the purpose of parameter estimation. Numerical simulation is carried out and results are presented to demonstrate effectiveness of the proposed learning algorithms. The experiment results show that the proposed algorithm can provide accurate estimation for the error model. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
