Application of ridge regression analysis to optimum array processing

In this paper, the ridge regression. technique of Hoerl and Kennard is applied to array processing. Optimum solutions as described by Edelblute, et al. often result in a highly sensitive supergain condition. Thus in order to actually achieve the theoretical gain predicted, the excitation and position of each sensor must be controlled with extraordinary accuracy. We present a method of modifying the Wiener optimum solution based on the ridge trace in such a way as to alleviate supergain conditions with neligible reduction of array gain. We demonstrate this effect by examining the array gain and several measures of sensitivity.