本文述及一种识别非线性动力学系统的改进算法。由于以物理方法分析识别非线性项比较困难,故本文提出一种“有序的”正交回归技术以识别描述非线性恢复力的展开级数的系数。算法可以选出为代表恢复力所需的函数形式。为处理在识别模型输入有噪声污染的问题,处理系数识别的顺序被排定,这样的处理有助于提高本方法的、特别是在小的信号幅度情况正气抗干扰能力。 This paper deals with an improved algorithm for identifying nonlinear dynamic systems. Since it is difficult to analyze and identify nonlinear items by physical methods, this paper presents an “ordered” orthogonal regression technique to identify the coefficients of the expansion series that describe nonlinear restorative forces. The algorithm can be chosen as the function form required to represent resilience. In order to deal with the problem of noise pollution input to the identification model, the order of processing coefficient identification is scheduled. Such processing helps to improve the positive air-immunity capability of this method, especially at small signal amplitudes.