A new method for searching the optimal step size of NLMS algorithm in intelligent antennas arrays based on fuzzy logic and artificial neural networks

This paper proposes a new approach based on information classification algorithms, fuzzy logic and neural networks for generating the membership functions to be used for searching the optimal step size of the normalized least mean square adaptive algorithm, when it is applied in adaptive linear antenna arrays. To this end a multilayer perceptron neural network, which is trained using the back propagation algorithm and the information obtained from several values estimated from the cost function of NLMS algorithm as well as the signal to interference noise ratio, generates the membership functions. Experimental results obtained by computer simulation show that the proposed approach provides fairly good estimation of the membership functions of the fuzzy logic stage used to obtain a near optimal step size. Evaluation results also show the desirable convergence properties of proposed approach when used to optimize adaptive antenna arrays.