Transient performance study of gradient algorithm of antenna array adaptation

Abstract

The paper presents a study into transient performance of the original gradient algorithm for the criterion of minimal noise dispersion. Synthesized during previous research, original gradient algorithm for the criterion of minimal dispersion shows an advantage in that it does not require formation of a reference signal for the adaptive antenna array adjustment. It is stressed, that to fully assess the possibility of practical implementation of the designed algorithm, it is necessary to carry out its comparative analysis with classical gradient algorithm of the steepest descent for root mean square error criterion. Comparison for the two algorithms is done for four interference situations by the way of computer simulation. While analyzing simulation results, it is found that adaptation of the both algorithms to a large extent is determined by the interfering signals correlation matrix parameters of antenna array, namely its eigenvalues spread. During transient performance estimation, it is revealed a tendency that the curve of adaptation for the original gradient algorithm for minimal noise dispersion criterion consists from some distinct segments with different steepness. It is postulated that these segments are determined by the number and the magnitudes of the eigenvalues of the interfering signals correlation matrix of the antenna array. The main conclusion is the one that the original gradient algorithm for minimal noise dispersion criterion and the classical gradient algorithm of the steepest descent for root mean square error criterion demonstrated pretty equal performances in the different interference environments. From the computer simulation, it is stressed that original gradient algorithm for antenna array adaptation for the noise minimum dispersion criterion can be used instead of the gradient algorithm of the steepest descent for the root mean square error criterion, especially in those applications where formation of the reference signal is impossible or problematic.