摘要
本文提出一种在频率域内直接估计随机动态系统的连续时间模型参数的快速迭代算法。该算法将时域的广义误差法推广到频率域。在估计模型参数的过程中,所得到的是关于参数的线性正则方程组。避免了非线性估计所带来的困难,附录中证明了所得到的参数估计量是一致的,数字仿真的例子也表明本文所提出的算法是有效的。
In this paper, a direct procedure for estimating the parameters of a continuous-time model of a dynamical stochastic system is presented. It is a fast interative algorithm in frequency domain. The generalized error method in time domain is extended to frequency domain by this method. A set of linear normal equations of the parameters can be obtained by optimizing the cost function, and the difficulties associated with nonlinear parameter estimation can be avoided. It is shown in the appendix that the parameter estimators obtained are consistent. The results of digital simulation of a continuous system demonstrate that the algorithm suggested is effective.
出处
《控制理论与应用》
EI
CAS
1986年第4期30-41,共12页
Control Theory & Applications
