Modifying the generalized singular value decomposition with application in direction-of-arrival finding (Q1387252)

The classes of updating and downdating algorithms for modifying the generalized singular value decomposition (GSVD) of a matrix pair are developed. The method based on CS decomposition is more accurate but computationally more expensive than the method based on the generalized eigenvalue decomposition. In both cases the updating and downdating problems can be reduced to a rank-one SVD updating problem. Applications in direction-of-arrival finding are derived. Numerical experiments are carried out to show the tracking ability of the algorithms.