The identification theory and method of fractional system is an important research direction in the field of fractional-order control which has drawn much research attention recently. However, the relevant study rarely involved the structure and order of fractional-order systems. Firstly, the colored noise model of linear fractional-order system is discussed. Then, the condensed information matrix whose information vectors have shift structure character is constructed, and the theoretical analysis and its proof about the condensed information matrix factorization and transform are also given. On this basis, an identification algorithm by transforming the condensed information matrix alternately and extracting relevant information from this matrix is proposed, and the decision criteria for the optimal estimation model structure and order are discussed, so the structure and order and parameters of the fractional-order system time domain model can be identificated simultaneously. The simulation study indicates that this method can obtain satisfactory identification results, and also has better ability of noise suppression. Not only the structure and order of the system model but also those of the noise model can be identified accurately.%在愈来愈被关注的分数阶控制研究中,分数阶系统辨识的理论与方法是一个重要方向.目前相关研究极少涉及分数阶系统的结构和阶次辨识.首先讨论了分数阶线性SISO系统辨识的有色噪声模型,然后构造了具有“移位性质”的信息向量和信息压缩矩阵,并给出了信息压缩矩阵分解变换的理论分析及其证明；在此基础上,提出了一种利用信息压缩矩阵交替变换提取相关信息的算法,论述了最优估计模型结构与阶次的判定准则,从而同时辨识分数阶系统时域模型的结构、阶次与参数;仿真研究表明,本文方法能够获得满意的辨识结果,具有良好的抑制噪声干扰能力；不仅能够准确地辨识系统模型的结构与阶次,而且能够辨识噪声模型的结构与阶次.
展开▼
