MIMO identification of parametric models for ultrasonic reflection and transmission experiments

This paper presents a multiple input multiple output (MIMO) identification approach, combining both reflection and transmission data from experiments performed on a linear viscoelastic material at normal incidence. The parametric modeling of the wave propagation, the calibration of the measurement setup and the estimation of the model parameters are elaborated. The broadband absorption-dispersion functions are incorporated in the wave propagation model by means of a rational transfer function. It is shown that the redundancy in the MIMO identification approach (both reflection and transmission) leads to an improved estimate of the material properties with respect to reflection or transmission identification only