A linear filtering approach for direction-of-arrival estimation of coherent signals

The problem of estimating the direction of arrival from coherent sensor array data and the estimation of frequencies of superimposed exponential signals from a time series are considered. By exploring the special structure of the problem, a linear filtering approach is proposed to generate multiple snapshots of data which are not coherent. It is shown that the method is a generalization of the spatial smoothing approach described by J.E. Evans et al. (1981) which is shown to be optimal under certain conditions.<>