Subspace-based identification of infinite-dimensional multivariable systems from frequency-response data (Q1923073)

An identification algorithm which identifies low complexity models of infinite-dimensional systems from equidistant frequency-response data is presented. This new algorithm is a combination of the Fourier transform technique with the recent subspaces techniques. Given noise free data, finite-dimensional systems are exactly retrieved by the algorithm. When noise is present it is shown that identified models strongly converge to the balanced truncation of the identified model if the measurement errors are covariance bounded. Several conditions are derived on consistency and two numerical examples are presented.