Large modular structures for adaptive beamforming and the Gram-Schmidt preprocessor

Use of a vector space approach to derive modular structure for constrained minimum variance beamforming removes complex algebra and reinforces the concept of a link between modular structure and the Gram-Schmidt preprocessor. The factorization of the orthogonal projection operator in Hilbert space yields the modular structure. The Gram-Schmidt preprocessor is a special form of modular decomposition.