A Superresolution Method of ARMA Spectral Estimation

摘要

Recently, a method for generating an ARMA spectral estimator model which possessed superresolution performance was developed. This method entailed minimizing a weighted quadratic functional of a set of basic error terms. An issue which remained to be resolved at that time was the selection of the weighting matrix that characterized the functional being minimized. A weighting matrix selection procedure has recently been developed and is reported. The autoregressive parameters are found in this procedure by minimizing a weighted sum of squares of zero mean basic error terms. The new weight selection is chosen to provide more heavy weighting to those terms in the sum which possess lower variances. Empirical evidence indicates that this new weight selections provides superior spectral estimation performance when compared to the original (N-m) to the 4th power weight selection.