Low Complexity Partially Adaptive Antenna Array Based on Recursive Least Square Algorithm

Abstract

The adaptive array with a rectangular aperture, whose antennas are placed in the nodes of the rectangular grid, is considered in the paper. The partial adaptability is provided via the base-band combination of the signals in the rows and in the columns of the array. Opposite to the hybrid arrays, based on the subarray output signals combination, the arrays based on the row and the column signal combination, do not suffer on the grating lobes phenomenon, which restricts the ability of the cancellation of the interferences, whose source spatial locations are coincided with those of the grating lobes. A version of the Matrix Inversion Lemma Recursive Least Squares adaptive algorithm for the partially adaptive narrow-band array weight calculation is presented. The Householder transform or QR decomposition adaptive algorithms can be used as well. The algorithm arithmetic complexity does not depend on the number of the array antennas, but depends on the array columns number and the rows number. The simulation results demonstrate that the steady-state performance of the adaptive array with the partial adaptation is the same as that of the adaptive array with the full adaptation if the number of the received signal sources does not exceed the number of the partially adaptive array weights. The dynamic behavior of the both arrays in the terms of the radiation pattern values in the directions of the received signal sources is differed only at the initial stage of the transient response. The equivalent weights of the full array used for the radiation pattern calculations are reconstructed via the combination of the weights of the crossed rows and the crossed columns of the partially adaptive array. The considered partial adaptation can be used in the large rectangular antenna arrays.