<title>Penalty function method for sidelobe control in least-squares adaptive beam forming</title>

Abstract

The weights required to maximize the SNR for array processing are well known and used extensively. However, in realistic situations it is often desirable to consider other factors. A common problem is the reduction of sidelobe jitter and sidelobe suppression while simultaneously maximizing SNR. In this paper we discuss a penalty function approach which can be used to satisfy all of these requirements by constraining a beampattern to lie arbitrarily close to a desired or quiescent beampattern and also satisfy a conventional look direction gain constraint. By a simple transformation it is possible to show how, in certain situations of physical interest, the method is equivalent to constraining the weight vector itself. We show how this method can be used for linear array and how the performance is affected by varying a user defined parameter. We conclude by showing how the method can envisaged as being equivalent to both a subspace projection technique and to diagonal loading with an extra quiescent constraint.