Numerically robust transfer function modeling from noisy frequency domain data

@article{Bultheel2005NumericallyRT,
  title={Numerically robust transfer function modeling from noisy frequency domain data},
  author={Adhemar Bultheel and Marc Van Barel and Yves Rolain and Rik Pintelon},
  journal={IEEE Transactions on Automatic Control},
  year={2005},
  volume={50},
  pages={1835-1839}
}
Using vector orthogonal polynomials as basis functions for the representation of the rational form of a linear time invariant system, in frequency domain identification problems, it is shown that the notorious numerical ill conditioning of these maximum likelihood problems can be overcome completely. For the identification of high-order (100/100) systems operating over a wide frequency band, or even in the situation of over- or undermodeling, condition numbers less than ten are reported for… 
Robust system identification of continuous-time model from frequency response function data
This paper addresses the numerical conditioning problem that arises in the continuous-time system identification case. To solve this problem, a new frequency domain weighted least squares estimator
Numerically Reliable Frequency-Domain Estimation of Transfer Functions: A Computationally Efficient Methodology
Abstract Parametric identification of lightweight motion systems requires solving large weighted least-squares problems. The numerical conditioning of such problems, which determines the solution
Frequency localising basis functions for wide-band system identification: A condition number bound for Output Error systems
Identifying the parameters of a system possessing a large dynamic range presents a number of difficulties for a least squares type algorithm. Foremost, is that of obtaining reliable parameter
Bi-Orthonormal Polynomial Basis Function Framework With Applications in System Identification
TLDR
The proposed method is built on the introduction of bi-orthonormal polynomials with respect to a data-dependent bi-linear form that generalises the conventional inner product and allows for asymmetric and indefinite problems.
A Frequency Localizing Basis Function-Based IV Method for Wideband System Identification
TLDR
The purpose of this paper is to present a numerically reliable simplified refined IV method able to identify (highly) resonant systems over a large bandwidth, from frequency-domain data.
Numerically Reliable Identification of Fast Sampled Systems: A Novel δ-Domain Data-Dependent Orthonormal Polynomial Approach
TLDR
A data-dependent orthonormal polynomial approach is developed for systems parametrized in the δ -domain that effectively addresses both the numerical conditioning issues encountered in frequency-domain system identification and the inherent numerical round-off problems of fast-sampled systems in the common Z-domain description.
Frequency-domain instrumental variable based method for wide band system identification
TLDR
An iterative IV scheme coupled with an appropriate frequency localising basis function is developed to provide consistent results in the case of highly resonant wide-band systems.
...
...

References

SHOWING 1-10 OF 25 REFERENCES
Best Conditioned Parametric Identification of Transfer Function Models in the Frequency Domain
In this paper, it is shown that rational transfer function models based on orthogonal Forsythe polynomials minimize the condition number of the Jacobian of estimators in a least-squares framework. As
On the use of orthogonal polynomials in high order frequency domain system identification and its application to modal parameter estimation
In this paper a maximum likelihood frequency domain identification technique is presented for the parametric transfer function of very high order linear time invariant systems. Through the use of
Generating robust starting values for frequency-domain transfer function estimation
Parametric identification of transfer functions in the frequency domain-a survey
TLDR
This paper gives a survey of frequency domain identification methods for rational transfer functions in the Laplace (s) or z-domain through a study of the (equivalent) cost functions.
Parametric and nonparametric identification of linear systems in the presence of nonlinear distortions-a frequency domain approach
TLDR
A related linear dynamic system (RLDS) approximation to the nonlinear system (NLS) is defined, and it is shown that the differences between the NLS and the RLDS can be modeled as stochastic variables with known properties.
Analyses, development and applications of TLS algorithms in frequency domain system identification
TLDR
An overview of frequency domain total least squares (TLS) estimators for rational transfer function models of linear time-invariant multivariable systems is given and the asymptotic behavior of the GTLS and BTLS estimators is studied.
On the frequency scaling in continuous-time modeling
TLDR
It is shown that the optimal frequency scaling also strongly depends on the estimation algorithm and that the median of the angular frequencies is a better compromise for improving the numerical stability than the arithmetic mean.
Transfer function synthesis as a ratio of two complex polynomials
Experimental frequency response data obtained from a linear dynamic system is processed to obtain the transfer function as a ratio of two frequency-dependent polynomials. The difference between the
...
...