线性多变量系统的辨识规范型及结构辨识算法

A canonical identification model for linear multivariable systems and its structural identification algorithm

从线性多变量系统状态空间规范型导出了系统的新输入输出辨识规范型，该模型直接与系统的输入输出数据相联系，且待辨识的参数较少，与现在的Ｇｕｉｄｏｒｚｉ辨识模型相比，大大减少了各子系统输出间的重迭（ｏｖｅｒｌａｐｐｉｎｇ）数目。给出了确定其结构指标的算法。数值实例表明了算法的有效性。

In this paper, a new input-output canonical identification model for linear multivariable systems is derived from its state-space canonical form. This canonical model with fewer parameters to be identified is directly related to the input-output of the system. Compared with the Guidorzi identification model, this canonical model greatly reduces the occurrence of overlapping in outputs of various subsystems. The algorithm for determining its structural indices is given. The numerical example shows that the algorithm works quite well.