Understanding Massive MIMO technology

To meet the demanding requirements laid down in the international specification for 5G, engineers adopted a fundamentally different approach to wireless network design. The increase in performance of 5G New Radio (NR) is based on a number of key technologies (including Massive MIMO and Beamforming) and the use of OFDM, a highly spectral-efficient modulation technique, and mmWave spectrum.

In this article, we' take a closer look at Massive MIMO, a key enabler of 5G, and its role in increasing data throughput and capacity.
 

What is Massive MIMO?

MIMO, (multiple-input, multiple-output) is a radio antenna technology which deploys multiple antennas at both the transmitter and receiver to increase the quality, throughput, and capacity of the radio link. MIMO uses techniques known as spatial diversity and spatial multiplexing to transmit independent and separately encoded data signals, known as "streams", reusing the same time period and frequency resource.

In multi-user MIMO (MU-MIMO), the transmitter simultaneously sends different streams to different users using the same time and frequency resource, thereby increasing the network capacity. Spectral efficiency and capacity can be improved by adding additional antennae to support more streams, up to the point where power sharing and interference between users result in diminishing gains and, eventually, losses. 

MIMO is used in many modern wireless and RF technologies, including Wi-Fi and Long-Term Evolution (LTE). 3GPP, the global organisation responsible for the definition of wireless standards, first specified MIMO for LTE in 2008, in its Release 8. This initial variant used two transmitters and two receivers, 2x2 MIMO, and subsequent increases in processing power have enabled the use of more simultaneous data streams in wireless networks with current 4G LTE networks using 4×4 MIMO.

The very short wavelengths at mmWave frequencies result in smaller antenna dimensions and for 5G NR, 3GPP has specified 32 antennas (32 x 32 MIMO) in Release 15, which will rise to 64 and more in future releases. This expansion in the size of MIMO antenna has led to the term Massive MIMO.

Massive MIMO techniques

Figure 1: A radio signal can take multiple paths between transmitter and receiver (Source: Qualcomm)

Massive MIMO is based upon the three key concepts of spatial diversity, spatial multiplexing, and beamforming. MIMO builds on the fact that a radio signal between transmitter and receiver is filtered by its environment, with reflections from buildings and other obstacles resulting in multiple signal paths (see Figure 1, right).

The various reflected signals will arrive at the receiving antenna with differing time delays, levels of attenuation and direction of travel. When multiple receive antennas are deployed, each antenna receives a slightly different version of the signal, which can be combined mathematically to improve the quality of the transmitted signal.

This technique is known as spatial diversity since the receiver antennas are spatially separated from each other. Spatial diversity is also achieved by transmitting the radio signal over multiple antennae, with each antenna, in some cases, sending modified versions of the signal.

Whilst spatial diversity increases the reliability of the radio link, spatial multiplexing increases the capacity of the radio link by using the multiple transition paths as additional channels for carrying data. Spatial multiplexing allows multiple, unique, streams of data to be sent between the transmitter and receiver, significantly increasing throughput and also enabling multiple network users to be supported by a single transmitter, hence the term MU-MIMO.

Beamforming uses advanced antenna technologies to focus a wireless signal in a specific direction, rather than broadcasting to a wide area (see Figure 2, below).

Figure 2: Beamforming focuses a wireless signal in a specific direction (Source: Everything RF)

Beamforming is another key wireless technique which works in unison with Massive MIMO to increase network throughput and capacity.

Beamforming uses advanced antenna technologies to focus the wireless signal in a specific direction, rather than broadcasting to a wide area. This technique reduces interference between beams directed in different directions, enabling the deployment of larger antenna arrays.

The large number of antennae in a Massive MIMO system enables 3D beamforming, which creates both horizontal and vertical beams toward users, increasing data rates (and capacity) for all users — particularly useful in urban areas with high-rise buildings.

Both the network and the connected mobile devices in MIMO systems must be tightly co-ordinated. Complex algorithms use spatial information obtained from a Channel State Information Reference Signal, (CSI-RS) to enable the base station to communicate with multiple devices concurrently and independently. The CSI-RS is a type of pilot signal sent out by the base station to the UE, which enables the UE to calculate the Channel State Information (CSI) and report it back to the base station.

The CSI describes how the signal propagates from transmitter to receiver and includes information on how that signal suffers from effects such as scatter, fade and power decay over distance. To recover the transmitted data-stream at the receiver, the MIMO system decoder must perform a considerable amount of signal processing, using the CSI to represent the channel transfer function in matrix form (see Figure 3, below).

Figure 3: Channel State Information used to characterise a Massive MIMO system (Source: Analog Devices)

The channel transfer matrix is defined as:

[R] = [H] x [T]

Where [R] is the series of signals received at the various antennae in the MIMO array, [H] represents the properties of each signal path, and [T] the various data-streams being transmitted across the network.

The decoder constructs the channel transfer matrix by estimating the individual channel properties, h11, h12, etc. from the CSI. The individual data streams are then reconstructed by multiplying the received signal by the inverse of the transfer matrix:

[T] = [H]-1 x [R]

Estimating the individual channel properties and computing the inverse channel matrix is computationally intensive and can add significant overhead to the network, particularly as the number of antennas grows.

The above description is somewhat simplified as there are actually various techniques for acquiring and calculating the CSI which depend upon factors such as the multiplexing techniques used, (TDD or FDD), the signal frequencies, and the amount of movement of the UE. This area is the subject of much ongoing research into how advanced techniques such as neural networking can enhance the reliability and accuracy of Massive MIMO.
 

The benefits of Massive MIMO

As a key building block of 5G NR, massive MIMO brings multiple benefits to both network operators and end users. The technology significantly improves spectral efficiency, delivering more network capacity for the same amount of spectrum, thereby enabling operators to maximise their investments in this expensive resource.

As 5G networks are rolled out they will depend heavily on network densification in order to deliver the required data rates and to support the high number of connections, particularly in urban areas. Massive MIMO, in conjunction with beamforming technology enables highly targeted use of spectrum, removing current performance bottlenecks, supporting a larger number of users in the cell, and improving end-user experience in densely populated areas.

Other potential benefits include higher connection reliability along with increased resistance to interference and intentional jamming, due to the increased number of signal paths. Massive MIMO networks will also be more responsive to devices transmitting at higher frequencies, which will improve coverage, particularly indoors.
 

The future of Massive MIMO

As 5G networks roll out, the use of Massive MIMO will expand, with ever larger antenna arrays becoming feasible as the technology and 3GPP specifications evolve. mmWave is the key to 5G performance and capacity, and massive MIMO arrays which can operate at these frequencies will soon become mainstream. NEC, for example, has developed a prototype 24-antenna array capable of operating at 28 GHz, and commercial Massive MIMO systems, with 64 arrays or more, will soon be mainstream at both sub-6 GHz and mmWave frequencies.

These deployments will be facilitated by the parallel evolution of Advanced Antenna Systems (AAS), which integrate the antenna arrays with the associated RF transmission hardware and software, as well as the signal processing capability required by beamforming and MIMO. As mmWave shrinks the size of the antennae and also the electronic components, these AAS will become smaller, playing a key role in network densification, and being deployed to provide 5G coverage in indoor locations.