Cascaded adaptive arrays

One of the problems with the least mean square (LMS) adaptive array is limited dynamic range caused by eigenvalue spread of the input covariance matrix, A solution is to process the LMS input signals prior to the LMS processor in power inversion arrays. The power inversion arrays attenuate the more powerful signals, thereby reducing the LMS eigenvalue spread and extending the overall dynamic range. Systems of power inversion arrays followed by an LMS processor are termed cascaded adaptive arrays. This dissertation analyzes cascaded adaptive arrays. It is shown that many of these cascaded arrays possess the same steady state output as an independent LMS array. This is desireable because the LMS array maximizes signal to interference plus noise ratio in the output. Computer simulation of the transient response of cascaded arrays is also presented for simple cascaded arrays. The results are generalized to multi-element arrays. It is shown that it is possible in some cases to eliminate phase distortion of the desired signal caused by the adaption transient.