Transfer function identification from phase response data

This paper introduces an improved procedure for the identification of a transfer function from phase angle data with prescribed frequency variation. It is shown how a transfer function can be identified from phase response data samples, by fitting a normalized function with constant magnitude using the vector fitting algorithm. The presented approach is numerically robust and leads to more accurate results than conventional approaches.