Improved estimation of SEM parameters from multiple observations

Abstract

A two-step technique for the estimation of singularity expansion method (SEM) parameters is described. The first step uses the iterative preprocessing algorithm (IPA) to estimate poles from noisy transient data. Errors in the estimated poles are reduced when the IPA is extended and applied to multiple measurements. For the second step, an iterative scheme for estimating the remaining SEM parameters is proposed. The results provide a way to estimate the SEM description of a scatterer from multiple measurements taken at different locations and with different directions of incidence. For large-to-moderate signal-to-noise ratios (SNRs) and Gaussian noise, these estimates meet the Cramer-Rao lower bound and thus have the minimum variance of any unbiased estimator. >