Neural network approach to diagnose phase shifter faults of antenna arrays

The diagnosis of faulty phase shifter in a uniform linear phased array antenna using a new method is presented. For parallel feeding of antenna elements each element has a separate phase shifter. The phase shifter for any particular element may fail due to failure in drive electronics. The failures of phase shifter are called phase shifter faults. In this work, an artificial neural network approach is adopted to diagnose phase shifter faults. A linear array of 21 elements with uniform spacing and uniform excitation with a progressive phase shift of ¿ is considered. A feed forward back propagation algorithm is used to train a neural network with a deviation radiation pattern which is the difference between the measured radiation pattern of array with normal phase shifters and degraded radiation pattern of array with a faulty phase shifter. The network thus trained predicted the number of the antenna element with faulty phase shifter with a high success rate. This is illustrated in a confusion matrix.