Compressive Sensing in Massive MIMO Array Testing: A Practical Guide

Abstract

Compressive sensing (CS) has been deployed in a variety of fields including wideband spectrum sensing, active user detection, and antenna arrays. In massive multiple input multiple output (MIMO) arrays, CS has been applied to reduce the number of measurements required to verify the arrays excitation. To date, the literature has concentrated on the various methods of achieving CS and applying them to both linear and 2-D arrays, and aimed at detecting fully failed elements in an array, offering simple pass/fail testing. All follow the general approach of creating the sparsity needed for CS by subtracting the measured far field or near field of the test array from that of a “gold standard” array measured under exactly identical conditions. This article extended this work to the need for rapid but accurate reconstruction of element excitation in a production testing environment for massive MIMO arrays. The aim is to demonstrate that CS can offer accurate reconstruction of array excitation. Particularly, the work addresses the issues of optimal sampling, measurement noise, accuracy of faulty element detection, effects of beam scanning, and physical alignment of the gold standard array with the test array. We have restricted ourselves to considering production standard arrays with failure rates up to around 5% and conclude with a set of proposed modifications to the basic CS process as applied to array excitations that achieve a near 20 dB improvement in the accuracy of the reconstructed array excitation offering mean square errors (MSEs) near to −40 dB, with a sampling strategy of just 1.4% of the Nyquist rate. This is achieved with the number of measurements to array element size ratio of approximately 0.2.