A generalized approach to high-resolution array processing

Many techniques have been proposed for the high-resolution processing of an array having an arbitrary geometry. We show that many high-resolution techniques can be represented using a common formulation based on the eigenvectors and eigenvalues of the correlation matrix. This formulation suggests a family of high-resolution spectral estimators governed by a single parameter that can be selected to give continuous adjustment of the spectral resolution. Setting the resolution parameter also affects the bearing-estimation accuracy, and we investigate the trade-off between resolution and accuracy using a simulated two-dimensional array.