论文部分内容阅读
本文提出一种基于奇异值分解（ＳＶＤ）的递推最小二乘辨识新方法，该方法不仅有很好的收敛性和数值稳定性，而且在系统的输入信号不满足持续激励的充分必要条件下，仍能得到系统参数的无偏估计，仿真计算结果证明了本文方法的有效性和优越性．
In this paper, a new recursive least square method based on Singular Value Decomposition (SVD) is proposed. This method not only has good convergence and numerical stability but also satisfies the necessary and sufficient conditions for the continuous excitation of the input signal , The unbiased estimation of system parameters can still be obtained. The simulation results show the effectiveness and superiority of the proposed method.