Título: Identification of linear process dynamics by means of integral equation models.
Autores: Lee, Howard Chong.
Fecha: 1964
Publicador: McGill University - MCGILL
Fuente:
Tipo: Electronic Thesis or Dissertation
Tema: Electrical Engineering.
Descripción: A method has been developed for identifying linear processes having one input and one output in the presence of noise, with the assumption that some a priori knowledge of the process dynamics is available. The method fits an integral equation model to the observed input and output of the process using the principle of least squares. The model parameters together with the initial conditions of the output are determined simultaneously without requiring a test signal. This method has been applied to identify various types of process dynamics, including three hand-drawn unit-impulse responses, by simulation on a digital computer. During this study both the a priori knowledge and the output signal-to-noise ratio of the process were varied. A process con be accurately identified within a fraction of its settling time for a reasonable signal-to-noise ratio at the output. For the case of using a model with more parameters than necessary, the estimates of the unwanted parameters are approximately zero, provided that (a) the noise level is low, and (b) all these unwanted parameters are in either the numerator or the denominator of the model transfer function. Also, a relatively fast time-varying process has been successfully identified in the presence of noise. The results indicate that this method is particularly effective when low order models are selected. [...]
Idioma: en