SMART ANTENNA ADAPTIVE BEAM FORMING BASE ON NEURAL NETWORK WITH DIFFERENT TRAINING ALGORITHIMS

Abstract

In this paper, an artificial Feed Forward Neural Network(FFNN) is used for smart antenna adaptive beamforming. Neural network is used to calculate the optimum weights that adapt the radiation pattern of the antenna by directing multiple narrow beams toward the desired users and nulling interference or unwanted users. Supervised learning algorithms were used to train the FFNN that used as adaptive beam former. The FFNN was trained using Levenberg-Marquardt (LM), Resilient Back propagation(RP) , Gradient Descent With Adaptive Learning Rate Backpropagation (gda) and Gradient descent with momentum and adaptive learning rate back propagation (gdx).The simulation results are applied for uniform linear array with five antenna element and the spacing between element equal to half wavelength. The results show that the best system performance can be obtain when the network trained by Levenberg-Marquardt (LM) algorithm.