| 随机状态模型参数辨识的多步算法 |
| |
| 引用本文： | 袁爱进,尹迅海,吴文芝.随机状态模型参数辨识的多步算法[J].大连铁道学院学报,1988(3). |
| |
| 作者姓名： | 袁爱进 尹迅海 吴文芝 |
| |
| 作者单位： | 大连铁道学院自动化教研室，大连铁道学院自动化教研室，中国人民警官大学六系 |
| |
| 摘 要： | 基于解耦辨识和多步计算思想,本文提出了一种状态空间模型参数辨识的多步算法.计算法包括二步加权最小二乘法、一步输出信息序列的修正和一步自适应 Kalman滤波过程,具有全局收敛且对待辨识参数初始估值设置不敏感等特点.仿真结果表明,在受控系统承受测量噪声和过程噪声的情况下,该算法对线性状态空间模型的参数辨识是十分有效的.
|
| 关 键 词： | 辨识 参数估计 最小二乘法 逼近 自适应滤波器/卡尔曼滤渡 |
A Multi-Step Algorithm for the Parameter Identification of Stochastic State Space Models |
| |
| yuan Aijin yin Xunhai Wu Wenzhi.A Multi-Step Algorithm for the Parameter Identification of Stochastic State Space Models[J].Journal of Dalian Railway Institute,1988(3). |
| |
| Authors: | yuan Aijin yin Xunhai Wu Wenzhi |
| |
| Institution: | yuan Aijin yin Xunhai Wu Wenzhi |
| |
| Abstract: | Based on the decoupling of system parameters and states estimation and concept of mul- ti-step procedure,a multi-step algorithm for parameter identification of stochastic state space models is proposed.The algorithm is constituted by two stages of Weighting Least-square,a refining on output residual sequence and an adaptive Kalman filtering.It is overall convergent and insensitive to the initial estimates of the system parameters.Simulation results show that the multi-step algorithm is sufficiently effective for parameter identification of linear stochastic state space models which are Contaminated by the measurement and process noise. |
| |
| Keywords: | identification parameter estimation least squares method approximation adaptive lilter/kalman filtering |
| 本文献已被 CNKI 等数据库收录！ |
