The Effect of Fluctuations of Weighting Coefficients on the Statistical Characteristics of an Adaptive Antenna Array Tuned According to the Hebb Algorithm

摘要

The results of analyzing the statistical characteristics or the weighting vector of an adaptive antenna array tuned according to the Hebb algorithm, which is one of the classical learning algorithms used for tuning of artificial neural networks, are presented. The analysis is performed with consideration for fluctuations of the weighting vector. It is shown that fluctuations lead to distortions of the antenna-array output signal and to the mismatch effect (the presence of the residual power at the antenna output) that arises owing to fluctuations of weighting coefficients. In an adaptive antenna array tuned according to the Hebb algorithm, the overcompen-sation effect is not observed.