A low complexity 2-D adaptive array for separating uncorrelated sources
C.C. Ko · 2002
This paper describes a 2-D adaptive array for separating uncorrelated sources. A matrix preprocessing beamformer, which will result in outputs from individual sources in the steady state, is used. The beamformer is updated periodically using the most recently available source location estimates. These estimates are in turn refined by using the beamformer outputs to adaptively eliminate correlated components in a reference element of the array. Although its initial convergence may be rather erratic, the adaptive algorithm has low O(M) complexity and has only 1 asymptotic time constant.