Hybrid design method for the partially adaptive arrays

A partially adaptive array is proposed that performs the degree-of-freedom reduction using two successive linear transformations, to create a "hybrid" transformation. The transformation that makes the first reduction is designed "offline" using an a priori distribution on the interference parameters. The transformation that makes the second reduction is designed "online" by estimating the correlation coefficients of the desired signal with the outputs of the first transformation, and choosing the outputs with the highest correlation coefficients to be adaptively weighted. We consider several hybrid designs that pair various offline and online methods. We find that the hybrid that pairs the offline power space method reported by the authors combined with the online modified Gram-Schmidt with correlation ordering (MGSCO) method of Yu and Lee (1994) outperforms the other methods we considered.