General formulation of a Prony based method for simultaneous identification of transfer functions and initial conditions

Previous results on transfer function identification using Prony signal analysis methods are extended. The class of allowed input signals is expanded: the signals can exhibit jump discontinuities and can be characterized by a finite number of eigenvalues between discontinuities. The advantage of this more general input form is that it can be tailored to excite system modes in frequency ranges of interest. In addition to generating system eigenvalues and transfer function residues, the approach can be used to obtain a feedthrough gain and initial condition residues. The approach makes maximum use of available input-output data during each phase of the solution. For the first phase of the solution, in which system eigenvalues are estimated, several alternative approaches are described. For the second phase of the solution, all available data are used in determining estimates of transfer function residues and initial condition residues. Methods of model order selection also are described, and a detailed example is given to illustrate the approach.<<ETX>>

[1]  L. Scharf,et al.  Statistical Signal Processing: Detection, Estimation, and Time Series Analysis , 1991 .

[2]  R. Kumaresan,et al.  Estimation of frequencies of multiple sinusoids: Making linear prediction perform like maximum likelihood , 1982, Proceedings of the IEEE.

[3]  S.M. Kay,et al.  Spectrum analysis—A modern perspective , 1981, Proceedings of the IEEE.

[4]  J. F. Hauer,et al.  Identifying linear reduced-order models for systems with arbitrary initial conditions using Prony signal analysis , 1992 .

[5]  Alan V. Oppenheim,et al.  Discrete-Time Signal Pro-cessing , 1989 .

[6]  Andrew J. Poggio,et al.  Evaluation of a Processing Technique for Transient Data , 1978, IEEE Transactions on Electromagnetic Compatibility.

[7]  L. Marple A new autoregressive spectrum analysis algorithm , 1980 .

[8]  Benjamin Friedlander,et al.  A modification of the Kumaresan-Tufts methods for estimating rational impulse responses , 1986, IEEE Trans. Acoust. Speech Signal Process..

[9]  Donald A. Pierre,et al.  An application of Prony methods in PSS design for multimachine systems , 1991 .

[10]  D. A. Pierre On the Simultaneous Identification of Transfer Functions and Initial Conditions , 1992, 1992 American Control Conference.

[11]  R. Kumaresan,et al.  Singular value decomposition and improved frequency estimation using linear prediction , 1982 .

[12]  L. Scharf,et al.  A Prony method for noisy data: Choosing the signal components and selecting the order in exponential signal models , 1984, Proceedings of the IEEE.

[13]  R. Kumaresan,et al.  Improved spectral resolution III: Efficient realization , 1980, Proceedings of the IEEE.

[14]  P. Barone Some practical remarks on the extended Prony's method of spectrum analysis , 1988 .

[15]  R. Kumaresan,et al.  Estimating the parameters of exponentially damped sinusoids and pole-zero modeling in noise , 1982 .