Parameter estimation of piecewise Hammerstein systems
Parameter estimation of piecewise Hammerstein systems
复制标题
DOI：
10.1177/0142331214531007
复制
发表时间：
2014-05
影响因子：
1.8
通讯作者：
Feng Wang;K. Xing;Xiaoping Xu
中科院分区：
文献类型：
--
作者：
Feng Wang;K. Xing;Xiaoping Xu
在非线性系统辨识中，系统通常被描述为一系列连接在一起的块。这种面向块的模型由静态非线性子系统和线性动态系统构成。本文研究了具有分段非线性的Hammerstein系统的参数估计问题。Hammerstein系统是一个面向块的模型，其中一个静态非线性块后面跟着一个线性动态系统。其基本思想如下。首先应用关键项分离技术，然后构造相应的辅助模型。因此，系统的辨识问题转化为参数空间上的非线性函数优化问题。再次，所有的参数的估计得到的建议粒子群优化算法。最后，通过与现有方法的比较，验证了该方法的有效性。此外，所提出的方法进一步推广到估计具有不连续非线性Hammerstein系统。
The system is often described as a series of blocks linked together in non-linear system identification. Such block-oriented models are built with static non-linear subsystems and linear dynamic systems. This paper deals with the parameter estimation of Hammerstein systems with piecewise non-linearities, which is a blocked-oriented model where a static non-linear blocking is followed by a linear dynamic system. The basic idea is as follows. The key term separation technique is applied initially, and then a corresponding auxiliary model is constructed. Hence, the identification problem of the system is converted to a non-linear function optimization problem over parameter space. Once again, the estimates of all the parameters are obtained by a proposed particle swarm optimization algorithm. Finally, compared with the existing methods, the simulation results confirm that the presented method is valid. Moreover, the presented method is further extended to estimate Hammerstein systems with discontinuity non-linearities.