Uncertainty of transfer function modelling using prior estimated noise models (Q1413930)

In problems involving transfer function modelling, the model errors are often the limiting factor. Under suitable assumptions concerning the disturbing noise and plant model errors, the authors provide approximate and easy-to-use analytic expressions for the asymptotic covariance matrix of a frequency domain estimator using prior estimated noise models. These analytic expressions show how the covariance matrix is affected by the disturbing noise level, the unmodelled dynamics and the nonlinear distortions. They also allow prediction of the order of magnitude of the actual uncertainty of the estimates, and reveal the relationship between transfer function modelling using prior estimated noise models and the classical prediction error framework.