Supporting Massive IoT Connectivity With Sectorized Massive MIMO

We explore the feasibility of supporting a huge number, K, of concurrent Internet of Things (IoT) connections, using a sectorized Massive MIMO base station (BS) equipped with M service antennas where K is much larger than M. Due to the limited length of the coherence interval, only <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\tau _{p}$ </tex-math></inline-formula> resources are available for uplink reference signals (RSs) that provide <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\tau _{p}$ </tex-math></inline-formula> orthogonal RSs. When K is much larger than <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\tau _{p}$ </tex-math></inline-formula>, the orthogonal uplink RSs must be reused or random non-orthogonal RSs can be utilized to accommodate the large number IoT connections. In this severe RS contamination scenario, we consider sectorization to enhance network performance. If we divide a circular cell into S sectors and using <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\lfloor M/S \rfloor $ </tex-math></inline-formula> directional antennas per sector, the effective array gain reduces to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\lfloor M/S \rfloor $ </tex-math></inline-formula>. However, this loss is more than compensated by the reduction of RS congestion, leading to a substantial increase in spectral efficiency (SE) as S increases. Given the large number of IoT devices, we increase the number of sectors to uncover some useful characteristics of the sectorization to support massive IoT connections with Massive MIMO. We derive a theoretical approximation of SE and propose a simple and effective scheme that determine the number of sectors for given performance requirements. Simulation results show that the proposed algorithm can achieve around 48% of SE improvement compared to the conventional scheme.