CARMA模型离线最小二乘迭代辨识方法

基于迭代最小二乘原理,提出了辨识CARMA模型和输出误差模型参数的最小迭代算法。两个最小二乘迭代算法分别比递推增广最小二乘算法和辅助模型递推算法具有更高的参数精度和具有很快的收敛速度。最小二乘迭代辨识的基本思想是:采用交互估计理论和递阶辨识原理,在每步迭代计算中,参数估计依赖于噪声估计,反过来噪声估计通过前一次迭代的参数估计计算,二者执行了一个递阶计算过程。最后用仿真例子验证了提出的算法。

Least-squares-iterative Identification Algorithms for CARMA Models

Based on the iterative least squares principle, two least-squares-iterative algorithms are developed for CARMA models and output error models. These two iterative algorithms use all measured input-output data at each iteration, and thus have highly accurate parameter estimation and faster convergence rates than the recursive extended least squares algorithm and auxiliary model based recursive least squares algorithm, respectively. The basic idea is to adopt the interactive estimation theory and hierarchical identification principle, the parameter estimates rely on the noise estimates, and the noise estimates are computed by the preceding parameter estimates. This performs a hierarchical computation processes. Finally, the simulation results indicate that the proposed algorithm can produce high accurate parameter estimation.