A Compressed Sensing Approach for Array Diagnosis From a Small Set of Near-Field Measurements

Abstract

A technique for array diagnosis using a small number of measured data acquired by a near-field system is proposed. The technique, inspired by some recent results in the field of compressed sensing, requires the preliminary measurement of a failure-free reference array. The linear system relating the difference between the field measured using the reference array and the field radiated by the array under test, and the difference between the coefficients of the reference and of the AUT array, is solved using a proper regularization procedure. Numerical examples confirm that the technique gives satisfactory results in terms of failure detection with a reduction in the number of data of two orders of magnitudes compared to standard back-propagation technique and of one order of magnitude compared to the number of elements of the array, provided that the number of fault elements is small. This result is relevant in practical applications, since the high cost of large array diagnosis in near-field facilities is mainly caused by the time required for the data acquisition. Accordingly, the technique is particularly suitable for routine testing of arrays.