Applicability of big data analytics to massive MIMO systems

Abstract

The evolving 5G standards promise green communications with enhanced data services and significant link reliability. Massive multi input multi output (MIMO) techniques are the driving force behind green communications, since they provide better energy efficiency with reduced transmit power. The massive data generated from such mobile communication systems, is a rich data source of great value. Procuring useful analytics from this precious resource, a big data aware 5G mobile communication system can be developed. A particular choice of big analytics brings in the concept of large random matrix models and single ring law. In this paper, first, big data analytics is performed in the context of a mobile user communicating to, either a massive MIMO or a massive MIMO orthogonal frequency division multiplexing (OFDM) system. Constructive insights such as transmitted (source) signal correlation analysis (attributed to certain network events), channel correlation analysis (attributed to user mobility) have been extracted. Ring law also has its roots in signal detection, which suggests that few other signal detection algorithms may be suitable candidates for signal/channel correlation analysis. Therefore, second, a proposed extension of an information theoretic criterion (ITC) based signal detection algorithm, for correlation analysis, is compared with ring law. Using massive MIMO and MIMO-OFDM system simulations, the said correlation analyses have confirmed the prevalence of ring law. Third, it is deduced that integrating big data analytics with massive MIMO system improves spectral efficiency.