Fast near-optimum resolution of coherent sources by line arrays

Several fast methods for resolving two coherent sources on the basis of spectrum analysis of time series are compared against the maximum likelihood estimation of bearing in terms of performance and computational effort. The methods examined include the Tufts and Kuaresan method, subarray methods, minimum eigenvalue methods, and a sparse sampling technique. It is shown that although a level of performance close to that of the maximum likelihood method cannot be achieved without incurring a similar work load the methods discussed here can provide the benefits of computational straightworwadness.