时滞和阶次未知的双率采样输出误差系统辨识

Identification of dual-rate sampled output-error systems with unknown time-delays and orders

针对时滞和阶次未知的双率采样输出误差系统,在有限采样数据条件下,提出一种辅助模型正交匹配追踪迭代辨识算法.首先,根据双率采样数据建立目标系统的辨识模型;其次,考虑到输入通道的时滞与系统阶次未知,采用过参数化方法,通过设置足够长的无噪输出数据和输入数据回归项,得到一个稀疏度为待辨识参数个数的稀疏系统;然后,结合辅助模型思想和压缩感知中的稀疏恢复方法实现参数向量和无噪输出的交互估计,即利用正交匹配追踪算法估计参数向量,再利用该估计值构建辅助模型计算无噪输出,并以此更新参数估计向量;最后,根据得到的参数向量结构计算系统的阶次与通道的时滞.仿真实验表明,所提出算法能够利用少量采样数据实现系统参数、时滞和阶次的高精度联合估计.

An orthogonal matching pursuit iterative identification algorithm is proposed for dual-rate sampled outputerror systems with unknown time-delays and orders based on finite number of sampled data.Firstly,the identification model of the target system is established based on the dual-rate sampled data.Secondly,considering the input time-delay and the system order are both unknown,by means of the overparameterization method and taking the regression items of the noise-free output data and that of the input data as sufficient length,a sparse system is derived and its sparsity is the number of parameters to be identified.Furthermore,the idea of auxiliary model and the sparse recovery method in compressed sensing are combined for interactive estimation of the parameter vector and the noise-free outputs,where the parameter vector is estimated by using the orthogonal matching pursuit algorithm,the auxiliary model is constructed based on which and then the noise-free outputs are calculated accordingly and applied in turn for updating the parameter vector.Finally,the system order and the input time-delay are calculated according to the structure of the obtained parameter vector.The simulation experiments verify that the proposed algorithm can provide accurate joint estimation of system parameters,time-delay and order based on small amounts of sampled data.