Principal and minor subspace computation with applications

Fast algorithms for computing signal subspace frequency or bearing estimates without eigendecomposition are described. These algorithms are based on the LR and the power methods for computing the eigendecomposition of matrices. Signal and noise subspaces are then utilized to develop high resolution methods such as MUSIC and ESPRIT for sinusoidal frequency and direction of arrival problems. A simple squaring procedure is suggested which provides significant computational saving in comparison with methods based on exact eigendecomposition. Simulations showing the performance of these methods are also presented