Adaptive antenna studies for HF communications

The treatment is confined to several variants of the least mean square (LMS) algorithm that use time averages as a means of improving the accuracy of the antenna weight adjustments. The techniques are described, and their performances are compared. Version 2, which applies time averaging to the error/input products before weight adjustment, is found to be simple to implement in either the digital or analog sense. Among its advantages are a reduction in the proportion of signal power required for the pilot carrier, a reduction in the effect of noise on the asymptotic values of the weights (and therefore the performance), and an improvement in the adaptation rate. Version 3, where the time averaging is applied to the input covariance matrix and only the most recent weights are used in forming the weight corrections, is found to be the most effective method.