Frequency Response Function Identification of LPV Systems: a Global Approach with Application to Mechanical Systems

Frequency Response Function (FRF) modeling of Linear Parameter Varying (LPV) systems facilitates analysis, controller design and parametric modeling of a large class systems, including position-dependent mechanical systems. The aim of this paper is to enable FRF identification of LPV systems using global experiments. This is achieved by developing an appropriate definition of the FRF for input-output LPV systems and by developing a method to compute a statistically optimal estimator of the FRF, which reduces to the Empirical Transfer Function Estimate (ETFE) for frozen scheduling. The developed method is successfully used to estimate a position-dependent FRF of a wide-format printer, confirming the potential of the approach.