BROAD-BAND LINEARLY CONSTRAINED ADAPTIVE ARRAY-PROCESSING WITH A PREPROCESSED GRAM-SCHMIDT ALGORITHM

This paper presents and investigates a new preprocessed Gram-Schmidt algorithm for broadband linearly constrained arrays using tapped delay line processing. It is proposed that before the conventional Gram-Schmidt scheme is employed to reject unwanted interferences, the element signals should first be transformed by using a simple preprocessor to satisfy the look direction constraint and to reduce the correlation due to finite bandwidth effects. With this preprocessing, the implementation complexity of the algorithm is reduced by a factor equal to the number of taps behind each array element. Furthermore, due to this size reduction, the new algorithm also becomes faster in its convergence behaviour.