作者: R. Pintelon , J. Schoukens , Y. Rolain
DOI: 10.1016/S0005-1098(00)00002-9
关键词: Value noise 、 Noise 、 Noise measurement 、 Gradient noise 、 Time domain 、 Applied mathematics 、 Frequency domain 、 Approximation error 、 Mathematics 、 Parametric statistics 、 Control theory
摘要: This paper treats the identification of continuous-time models using arbitrary band-limited excitation signals. A modeling approach is presented that has following two advantages: (1)asymptotically (the amount data tends to infinity) there no approximation error over complete frequency band from DC Nyquist, (2)it allows identify general parametric noise models. The key idea combine a plant model with discrete-time (=hybrid Box-Jenkins structure).