Adaptive Sidelobe Cancelling Using Complex-Valued Canonical Variables

A sidelobe canceler is an adaptive array configuration structured to enhance the reception of a desired signal corrupted by unwanted interference signals in applications such as radar, sonar, and communication systems. Several performance criteria, including the minimum mean square error (MMSE) and maximum likelihood, have been defined for adaptive array problems. A new performance criterion, referred to as maximum correlation (MC), is introduced herein in the context of sidelobe canceling. The MC criterion is based on the concept of canonical variables introduced by Hotelling in 1936 to defined a canonical relationship between two sets (or vectors) of real valued random variables. In this report Hotelling's formulation is extended to handle complex-valued random variables for the special case of a complex-valued scalar and a complex-valued vector. In addition, it is shown that the weight vector solution for the MC criterion is equivalent, up to a scale factor, to the MMSE weight vector.