摘要
本文在初始模型确定的条件下,按照缓变对象的特点,提出拟合效果较优的模型滚动辨识方法。由于采用自然的对象强迫输入与响应输出的确定性拟合方法,避免了统计方法必须拥有的大量样本数据的缺点,有利于模型在线辨识的实时性。最后,举例说明本方法的应用。
On the condition of an initial model being determined, a rolling identification method based on three different coincidence model, together with an optimizing principle is proposed in this paper, with which there is no need for a large amount of experimental data and a forced input signal from external system and thus the real-time performance of the model can be ensured.
出处
《自动化与仪器仪表》
1996年第4期5-7,共3页
Automation & Instrumentation