Weighted Acma

The analytical constant modulus algorithm (ACMA) is a deterministic array processing algorithm to separate sources based on their constant modulus. It has been derived without detailed regard to noise processing. In particular, the estimates of the beamformer are known to be asymptotically biased. In the present paper, we investigate this bias, and obtain a straightforward weighting scheme that will whiten the noise and remove the bias. This leads to improved performance for larger data sets.