Fast detection and location of failed array elements using the fast SVM algorithm

Abstract

With the increased proliferation of array antennas, for a myriad of wireless applications, the monitoring and correction of array element failure is becoming more critical. In this paper, a fast Support Vector Machine (SVM) classifier, based on low-degree polynomial kernel is developed, to expeditiously detect and locate the failed elements of an arbitrary analog array, while categorizing the level of failure, at the same time. This method will work on any analog array configuration, and the results for a linear array sample are illustrated to demonstrate its effectiveness.