基于子空间方法的非均匀多采样率系统辨识

针对非均匀多采样率系统的建模问题,根据因果关系,建立了非均匀多采样率系统的状态空间模型.对于含有提升变量的状态空间模型,提出基于子空间技术的辨识方法.首先,由系统的输入输出数据建立由Hankel矩阵组成的扩展状态空间方程;其次,利用斜交投影的原理,以及奇异值分解,通过子空间辨识算法确定增广观测矩阵和状态向量;最后,通过最小二乘方法确定模型的参数矩阵.该方法简单有效且对初值具有鲁棒性.仿真实例验证了方法的有效性.

Subspace-based method for identification of non-uniformly multirate sampling systems

According to the modeling of non-uniformly multirate sampling system, a state space model is derived due to the casual relationship. Subspace-based identification is developed for state space models, which have lifting variables. Firstly, an extended state space equation formed by input-output Hankel matrices is established. Then, the extended observability matrices and state vectors are obtained by subspace-based identification algorithm through the oblique projection and singular value decomposition. Lastly, the parameter matrices are determined using the least square algorithm. A simulation example is presented to illustrate the performance and robustness for initials of the proposed method.