| 引用本文: | 于淼,王友谊,郭戈,刘建昌.基于新息估计的递推闭环子空间辨识[J].控制理论与应用,2025,42(2):245~252.[点击复制] |
| YU Miao,WANG You-yi,GUO Ge,LIU Jian-chang.Recursive closed-loop subspace identification based on innovation estimation[J].Control Theory and Technology,2025,42(2):245~252.[点击复制] |
|
| 基于新息估计的递推闭环子空间辨识 |
| Recursive closed-loop subspace identification based on innovation estimation |
| 摘要点击 2749 全文点击 23 投稿时间：2023-12-23 修订日期：2024-11-23 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2024.30826 |
| 2025,42(2):245-252 |
| 中文关键词 子空间辨识 递推闭环辨识 新息估计 最小二乘逼近 |
| 英文关键词 subspace identification recursive closed-loop identification innovation estimation least squares approximations |
| 基金项目 国家自然科学基金项目(62003082, 61773106, 62173079, U1808205), 河北省自然科学基金项目(F2021501018), 河北省教育厅科学技术研究项目(ZD2022148)资助. |
|
| 中文摘要 |
| 本文针对闭环控制系统提出了一种基于新息估计的递推闭环子空间辨识方法. 首先, 通过新息估计减小闭环控制系统中过去噪声对未来输入的影响, 进而获得子空间矩阵的参数估计; 然后, 运用最小二乘的递推方法进一步降低噪声的影响, 从而实现下三角Toeplitz矩阵的参数估计; 最后, 采用Kung辨识算法对由估计参数构造的Hankel矩阵提取系统矩阵. 通过仿真实验验证了所提方法的有效性和优越性. |
| 英文摘要 |
| A recursive closed-loop subspace identification method based on innovation estimation is proposed for closedloop control systems. Firstly, the influence of past noise on future input in the closed-loop control system is eliminated by innovation estimation, and the estimation of subspace matrix parameters is obtained. Secondly, the recursive method of least squares is used to eliminate the influence of noise, which realizes the parameter estimation of the lower triangle Toeplitz matrix. Finally, based on the Kung’s realization algorithm, the system matrices are extracted from Hankel matrix constructed by the estimated parameters. The simulation results show the effectiveness and superiority of the proposed method. |
|
|
|
|
|