Neural network approach to diagnose faults in linear antenna array

A novel approach using artificial neural network (ANN) is proposed to identify the faulty elements present in a non uniform linear array. The input to the neural network is amplitude of radiation pattern and output of neural network is the location of faulty elements. In this work, ANN is implemented with two algorithms; radial basis function neural network (RBF) and probabilistic neural network and their performance is compared. The network is trained with some of the possible faulty radiation patterns and tested with various measurement errors. It is proved that the method gives a high success rate.