摘要
为了对带噪输出的控制系统进行盲辨识,提出了一种利用过采样法辨识控制系统传递函数的方法.通过过采样控制系统的输出信号,将原单输入单输出系统转换成为一个新的与原系统具有相同输出极点的单输入多输出系统.再估计出这个新的单输入多输出系统的传递函数,最后得到原单输入单输出系统的模型参数.仿真结果表明,与常用的高阶矩法相比,该方法具有不受系统结构影响,不需要计算样本数据的高阶统计矩,使用的样本数据量少,便于用计算机实现的优点.
In order to blind identify a control system whose output is corrupted by noise, a novel algorithm for identification control system transfer function is proposed based on an oversampling scheme. By oversampling the single input single output (SISO) system, the system is transformed into a single input multiple output (SIMO) one with common poles. With estimating the numerator of the new SIMO model transfer function first and the common denominator later, the parameter estimation of the original SISO model is obtained. The simulated result shows that, comparing with cumulants, this algorithm is not affected by different structure of the system, not needed to calculate the higherorder cumulants of the sampling, the sampling number requirement is small, and its realization is convenient by using computer.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2003年第2期171-174,共4页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(69974029).
