论文部分内容阅读
本文论述线性与非线性系统辨识方法。将辨识方法分为确定性与随机性两类,并据其对于辨识问题的利弊,扼要地讨论了有关参数、结构及状态估计方面的特点。本文注意到辨识的不同最终目的,同时也考虑到辨识方法的简便性。辨识方法最重要的发展趋势是解决那些没有解决的辨识问题及解决数字计算与实际应用的复杂性。对结构、参数或噪声的不定性影响和被分析系统的先验知识的可能应用,均作了研究。
This article discusses linear and nonlinear system identification methods. The identification methods are divided into two categories: deterministic and random, and according to their advantages and disadvantages for identifying problems, the characteristics of parameters, structure and state estimation are briefly discussed. This paper takes note of the different ultimate goals of identification, while also considering the simplicity of identification methods. The most important development trend of the identification method is to solve those unidentified identification problems and solve the complexity of digital calculation and practical application. The impact on the uncertainty of structure, parameters or noise and the possible application of prior knowledge of the system under analysis are studied.