摘要：AbstractA key step in experimental modeling of mechatronic systems is Frequency Response Function (FRF) identification. Applying these techniques to systems where measurement time is limited leads to a situation where the accuracy is deteriorated by transient dynamics. The aim of this paper is to develop a local parametric modeling technique that improves the identification accuracy of a range of systems by exploiting prior knowledge. The method is to impose a prior on the approximate locations of the system poles. This leads to better fit results and enables an accurate variance characterization. As a special case, traditional LPM is recovered.
