基于SVD的测量系统建模方法研究

Design of Indirectly Measuring Model Based on SVD

本文主要介绍了SVD方法的理论和性质,讨论了基于SVD的线性系统参数估计方法及其在间接测量系统建模中的应用,采用模块化、结构化矩阵处理函数,提出了一种以对观测数据矩阵进行SVD分解为基础的间接测量系统建模方法。并设计给出了一个数值分析实例,对文中方法的正确性和有效性进行了验证。仿真分析的结果表明该方法能够有效地去掉系统中的冗余信息,提高模型构建的准确度,尤其适合于大型测量系统的精确建模。

The paper introduced SVD （Singular Value Decomposition） method＇ s theory and property, discussed a novel method of linear system parameter estimation based on SVD and its application in the modeling of indirect measuring system. With the use of modular and structured matrix processing functions, a new method to model indirect measuring system was presented based on the SVD of observation data matrix. And a numerical analysis example was giyen, which verified the new approach＇s validity and correctness. The computer simulation results show this technique can discard system＇s redundant information and improve system model＇s accuracy effectively, and fits large measuring system＇s accurate modeling very well.