Amplitude-based spatial signal processing using high-order statistics. Part II

This paper continues the characterization and analysis of a spatial signal processing problem using only signal amplitude information. The results generalize the "signal separator" or "eigenvector sorting" algorithm used in some very low frequency (VLF) receiver systems by removing the requirement that the solution vectors be orthogonal. An algebraic characterization of the solution is formulated and used to facilitate a numerical search.<>