The Massive MIMO Innovation Pipeline

In a recent ABI Research executive foresight, we wrote about the “The Transformative Potential of Massive MIMO” and described the technology’s promises and challenges. In this companion executive foresight, we outline five encouraging research and development (R&D) areas in massive MIMO that point toward a future that promises prolific commercial-at-scale massive MIMO deployments as each of those challenges are tackled. We examine this massive MIMO innovation pipeline in the areas of extremely large aperture arrays (ELAAs), metamaterials, intelligent massive MIMO, architectures, and massive MIMO construction. While this is not intended to be a comprehensive list, ABI Research believes that these areas do in fact point toward the future of massive MIMO.

The theory behind massive MIMO has been well established, with many algorithms for communications, signal processing, and optimization, by academia. This has resulted in initial commercial deployments of massive MIMO antennas in the sub 6GHz bands for LTE and for 5G NR in both sub 6GHz and millimeter wave (mmWave) bands.

As 5G becomes a commercial reality, the industry envisions that digitally controllable massive MIMO antenna arrays will become pervasive, with conventional sites operating in the sub 6GHz bands equipped with massive MIMO antennas with 64 or more antenna elements per sector for wide area coverage and capacity. New basestations operating in mmWave spectrum with many more antenna elements will be deployed at rooftop or street level and indoors for local area coverage and capacity. This unprecedented ability to manage spatial resolution to give constructive interference at specific points in space and time, and resolve the ultra-fine details of received signals, is the key to increasing sector spectral efficiency and hence sector throughput for significant total cost of ownership reductions for the owners and operators of radio access networks.

In the following paragraphs, we sketch out five encouraging areas of research and development that are active in the academic and vendor community.

Extremely Large Aperture Arrays (ELAAs) depend on the fact that the spectral efficiency of the array increases as the number of antenna elements increases. What if the number of antenna elements were to reach hundreds or even thousands?

We can work around the practical limits of such scenarios by disaggregating the antenna elements from inside the antenna enclosure and placing them on the windows on the outside of a building, for instance. Instead of all the antenna elements in a single box, they are distributed over a substantially larger area. We can imagine a scenario where the user is surrounded by antennas instead of the more conventional scenario where the antenna is surrounded by users. This technique has been called cell-free massive MIMO, 5G coordinated multipoint (5G CoMP), or distributed massive MIMO. The benefits of this technique are that it will provide orders of magnitude higher throughput per sector when compared to conventional “compact” massive MIMO.

One way of doing this is to integrate the antennas and radio frequency front ends into cables attached to the façade of buildings and connected to a baseband either in proximity to the cable, in the basement of the building, or virtualized in a local data center. This type of cable is called a radio stripe, the concept of which imagines stringing up a series of radio stripes in much the same way as Christmas lights are used. In this way, many radio stripes share one common fronthaul connection to the baseband. Because there are many antenna elements in the ELAA, the antenna hardware can be constructed using low cost smartphone-grade hardware

Metamaterials also leverage the fact that the capacity of massive MIMO systems grows monotonically with the number of antenna elements. Unlike the ELAA, which distributes very large numbers of antenna elements across a large physical area, metamaterials promise a way to implement a very large number of antenna elements in a limited surface area in the form of a spatially continuous transmitting and receiving surface. This is known as holographic beamforming (HBF), where the array or electronically active surface interfaces with a single RF port that is connected to an electronically steerable radiating RF distribution network. The metamaterial consists of a controlled impedance substrate with a reverse biased varactor diode associated with each antenna element. By controlling the reverse bias voltage on each diode, the impedance of each of the antenna elements is altered, and beam steering becomes possible. The very high levels of spatial resolution mean that extremely low levels of transmit power can be used for each antenna element for very low-cost systems. The promise of holographic beamforming is that it can be integrated into many different types of surfaces including walls, glass, and even fabrics for very low cost while remaining aesthetically pleasing.

Intelligent Massive MIMO leverages the principles of artificial intelligence (AI) or machine learning (ML) to optimize the performance of a massive MIMO system. Examples of the use of AI/ML in massive MIMO include areas that are difficult to model with conventional approaches, but that can be calculated from data. These examples include, but are not limited to, channel estimation and feedback, massive MIMO power control, and user positioning. One opportunity intelligent massive MIMO opens up is the possibility of using very low-cost hardware, such as low precision analog-to-digital converters (ADCs), for dramatically reduced energy consumption and cost. The AI/ML algorithms would optimize the massive MIMO system for best performance and eliminate any impairments introduced by using low-cost/non-precision hardware. One of the key benefits of this approach, compared to ELAAs or holographic beamforming, is that less hardware is required. Also, these AI/ML algorithms could be used to dynamically alter the massive MIMO antenna array geometry and optimize the radio access network for best performance.

The architecture of a massive MIMO system can be digital, analog, or hybrid, with digital architecture as the best for flexibility and performance. Long viewed as high-power and high-cost, digital architectures have been avoided until now by the vendor community, which has preferred to build analog or hybrid systems for reduced time to market. However, at mmWave frequencies, fully digital massive MIMO systems become practical and feasible with antenna-on-chip or silicon antenna systems that leverage the Moore’s Law economies of scale of semiconductor technology. Other technologies promise to adaptively boost power amplifier efficiency according to demand, and the advent of low-cost ultra-wideband complementary metal-oxide-silicon (CMOS) RF integrated circuits and advanced techniques for beamforming, signal processing, and partitioning the beamforming processing between the antenna and baseband will also help reap the benefits of the economies of scale in component cost from the handset ecosystem. ABI Research expects that future massive MIMO implementations will become fully digital at both mmWave and sub 6 GHz frequencies.

Construction of massive MIMO antennas using waveguides has been shown to have between 3- and 10-times lower losses than substrate integrated waveguide antennas or microstrip antennas. Waveguide-based compact massive MIMO systems can be built to operate in the sub 6 GHz and mmWave spectrum with extremely low loss and high bandwidth and are compatible with fully digital beamforming architectures. Other benefits of this construction technique include high gain, compact form factor, and out-of-band interferer signal suppression. As a substrate-based solution, waveguide antennas are easily integrated as part of the monolithic microwave integrated circuit (MMIC) portion of the massive MIMO antenna.

We outlined only five of the R&D areas that have come to our attention for the massive MIMO antenna of the future and believe that this demonstrates the creativity and innovation in the massive MIMO ecosystem. With advancements in semiconductor technology, materials science, and communications, signal processing massive MIMO will emerge as the go-to technique for radio access network performance enhancement. Massive MIMO theory has long been known, but it is only now, with technology progression, that massive MIMO becomes feasible and practical for commercial-at-scale deployments in both sub 6 GHz and mmWave spectrum. We are about to witness a massive MIMO renaissance.