Data-Based Model for Wide Nulling Problem in Adaptive Digital Beamforming Antenna Array

Abstract

An adaptive digital beamforming (DBF) array is vital in advanced wireless systems. Beamforming problems in DBF arrays could be time-consuming and inflexible, as most of them are only solvable by optimization algorithms. In this letter, we propose an improvement to this issue in solving wide nulling problems of an adaptive DBF array by building a data-based wide nulling model using a powerful optimization algorithm—the Bat algorithm (BA) with the general regression neural network (GRNN). The BA efficiently generates the necessary full complex weights as training samples to train the model. The GRNN estimates the data in the gaps of training samples and gives an efficient and flexible estimation. A 32-element uniform linear array is used an example to demonstrate the proposed algorithm. Numerical experimental results show that the data-based model is functional with an acceptable performance for all test cases and achieves a much better time efficiency in comparison to using solely the BA in determining the complex weights to wide nulling problems.