Constrained iterative deconvolution applied to bearing and amplitude estimation of coherent sources

The authors introduce a constrained iterative deconvolution technique for coherent sources which applies two constraints at each iteration of the deconvolution: the estimated source energy must be nonnegative, and coherence may exist only within a predetermined separation in the spatial frequency (far-field) domain. They describe a modified constrained iterative deconvolution algorithm which can be applied to bearing and amplitude estimation. The technique has been shown, via the contraction mapping theorem, to converge to a nonunique solution, but one consistent with both the measured data and a priori knowledge of the sources