Abstract:

The plane wave reflection coefficient of an absorbing infinite plane can be estimated from pressure distributions on two lines parallel to the surface. The measured pressures are decomposed into incoming and outgoing plane waves in the wave-number domain after the Hankel transform is applied to the spatial data. In practice, only a finite number of pressure measurements can be made on the measurement lines. As a result, errors resulting from aliasing and truncation are induced in the wave-number spectra and subsequently the reflection coefficient. The aliasing results from insufficiently small sample spacing, and can be easily controlled. The aliasing results from insufficiently small sample spacing, and can be easily controlled. The truncation error results from the limited size of the measurement aperture, which may be difficult to expand for practical reasons. Therefore, extrapolation techniques based on parametric signal modeling have been developed to ``virtually'' expand the measurement aperture, thus minimizing truncation effects. However, some restrictions, particularly on spatial sampling rate, must be considered in order to obtain good estimates of the reflection coefficient. Prony series modeling is one procedure that can be used to model and extrapolate the spatial sound pressure data. Reflection coefficients estimated by using the Prony series approach will be used to illustrate this procedure.