In the previous article of this series, IHS investigated the key technologies which will enable Long Term Evolution - Advanced or LTE-A to break through the gigabit per second speed barrier. As the technology currently stands (3GPP Release 10), LTE-A is capable of aggregating a total spectral limit of 100MHz across 5 aggregated carriers to produce a theoretical data throughput of 750mbps downlink using conventional 64QAM modulation and 2 spatial layers (2x2 MIMO) per carrier. As discussed in the previous article, by leveraging higher order modulation of 256QAM encoding scheme instead of the conventional 64QAM, we can improve that downlink benchmark by as much as 33% in order to nudge the overall throughput to near the gigabit per second theoretical threshold. While higher order modulation is an important technology gain, it does not scale very effectively as incrementally higher order modulation schemes yields diminishing return in speed gains and optimizing the impact of 256 QAM requires high levels of channel quality. For better scaling of LTE-A performance, we need to exploit additional spatial multiplexing from the current limit of 2 layers per carrier to 4 layers (4x4 MIMO) per carrier. In this manner, speed gains scale linearly with the available spectrum and provide a more effective means to break through the gigabit per second speed barrier.
Long Term Evolution, as the LTE acronym is derived from, is an apt description of this constantly evolving wireless 4th generation (4G) technology platform. The goal of evolving existing 4G wireless technology is to produce an ever more efficient wireless data transport given limits of available spectrum, signaling schemes and the laws of physics. As advances in LTE and now LTE-A have pushed wireless transmission to the limits of Shannon-Hartley theorem, we must turn to other, non-RF (radio frequency) approaches to achieve gains. Beyond the dimensions of frequency, coding and time in LTE, we now turn our attention to leveraging the spatial dimension to further evolve and improve LTE performance. Conventional LTE handset and network communicate in tandem using a 2x2 MIMO configuration. This spatial multiplexing leverages 2 spatial layers or, in other words, signaling with 2 separate and distinct streams of data per carrier from the LTE network towers to the user device. This design corresponds to the dual antenna designs built into our smartphones today. However, in order to scale the performance of LTE-A, we can imagine extending this concept to 4 spatial streams or 4 antennas in order to double the capacity per carrier signal. This spatial multiplexing principle can be applied to LTE carrier aggregation thus a 4x4 MIMO across the maximum available 5 carriers yields a theoretical maximum of 20 spatial layers in current Release 10 LTE-A standard. Combined with higher order modulation, a staggering theoretical data throughput of 2Gbps can be realized representing a better-than order of magnitude improvement over initial 2x2 MIMO LTE (Release 8) standards while leveraging 5 times additional spectral resources . In real world network deployments, where spectrum availability is limited, the use of 4x4 MIMO represents a key LTE-A technology which will propel network performance and speed towards the 1 Gbps milestone.
Spatial Multiplexing; Theory and Real World Implementation of 4x4 MIMO
As with all LTE technology, the application of spatial multiplexing involves the coordinated efforts of both network architecture and handset RF front end design. Let’s take a deeper look into each area:
Network- LTE network deployments reflect significant capital expenditures (CAPEX) for network operators during the rollout of their new 4G infrastructure. With this long-term investment in mind, most of the currently deployed LTE base stations are already equipped with 4 antenna arrays at the LTE radio towers for future upgradeability to 4x4 MIMO. Some mobile operators, especially ones deploying TD-LTE, have gone one step further to future-proof their network RF componentry with 8 antenna arrays. While the vast majority of today’s LTE networks are operating in 2x2 MIMO mode which matches the capabilities of the existing LTE handsets, moving to 4x4 MIMO transmission can be achieve relatively easily through software update to the existing network equipment. However, given the economics of carrier CAPEX, network operators may have differing urgency to upgrade to 4x4 MIMO operations. For example, carriers that are rich in spectral holdings may opt to invest first in deploying additional LTE bands for carrier aggregation or better coverage while other carriers with limited spectral holdings may opt to jump to 4x4 MIMO sooner in order to improve network performance. Overall, the implementation of 4x4 MIMO on the network side is not so much limited by technology as it is by business economics and the criticality of available user devices (e.g. smartphones) capable of taking advantage of additional spatial streams.
Smartphone-Due to the limitations in physical size, smartphone RF designs have primarily relied on 2 physical antennas per LTE band. These tell-tale physical design limitations are clearly visible in today’s popular smartphones with a top and bottom antenna stripes strategically distanced for spatial diversity as well as to avoid signal attenuation from user hand grip. Evolving the RF design from 2x2 MIMO to 4x4 MIMO will require significant changes to the RF front end. Not only will smartphone designs need to accommodate 2 additional antennas per carrier but also absorb the cost of incorporating RF componentry along each new radio chain. Furthermore, not all LTE basebands on the market today are capable of taking advantage of 4x4 MIMO signaling therefore limiting the commercial availability of 4x4 MIMO smartphones.
RF Design Challenge; Adding more Antennas into a Limited Space
Clearly, the challenge to realize the performance gains from 4x4 MIMO will rest primarily at the device side of the overall LTE equation. Due to the design challenges listed earlier, smartphone RF designers will need to be creative in engineering around physical limitations of the handset. Here are two potential solutions RF designers can draw from in designing 4x4 MIMO smartphones:
- Antenna sharing with Wireless LAN RF; WiFi frequencies maps well to popular mid to high frequency LTE bands. These WLAN antennas can be utilized as common RF front end components at the cellular side of the smartphone. By reusing existing hardware, smartphone OEMs save on cost as well as physical space to execute a cellular 4x4 MIMO RF front end using compatible LTE spectrum. The challenge in implementing this antenna sharing scheme will be the close coordination of the cellular and WLAN baseband chips. This design approach favors chipset providers that produce products in both wireless categories as they can more effectively vertically integrate key functionality required for RF sharing. Qualcomm is clearly a leader in this arena.
- Co-location of antennas; Spatial multiplexing has traditionally relied on the principle spatially coherent antennas with physical separation. This configuration obviously runs into limitations in a small form factor of a smartphone. However, RF engineers have developed RF designs that circumvent this physical limitation by specially tuning co-located multiple antennas that have properties of high isolation and low correlation. By placing these antennas close to one another while still maintaining high antenna efficiency, RF designs can achieve the high spatial layer requirements of 4x4 MIMO in smartphones. One such implementation of this technology is called Isolated Mode Antenna (iMAT) from Skycross which allows a single antenna structure to behave like multiple antennas in a MIMO operations.
Given that frequency is inversely proportional to wavelength, not all LTE spectrums are ideal for 4x4 MIMO design. For example low frequency anchor spectrum used in the US market (700MHz) for greater coverage over larger distances necessitate the need for larger antenna structures relative to higher frequency LTE spectrum peers. Therefore, simply adding more low frequency antennas may not be attainable due to physical limits of devices. In practice, smartphones designed for those networks utilizing low band LTE frequencies will likely stay at a 2x2 MIMO configuration thus the average spatial streams across all aggregated spectrum will be lower than the theoretical maximum 4 spatial streams per carrier.
Drawing Parallels; Evolution of LTE to Gigabit LTE Advanced
In order to better illustrate the evolution LTE and the technologies highlighted in this and previous IHS articles on key LTE Advanced technologies, it would be useful to use a physical world analogy that can help to drive home the finer technical points made in this and earlier articles.
First, let’s think of LTE wireless communication in the analog of automobiles traveling in a highway system. The individual vehicles can be thought of as bits of information while the roads are analogous to the wireless spectrum available and decks of freeway represent the spatial layers created in multiplexing schemes. Thus the uni-directional communication between cellular tower antennas and the user device over one single LTE band can be represented in the graphic below.
From this initial conceptual framework, we can then extend that analogy to illustrate the current state of LTE Advanced deployments with the use of carrier aggregation and 256QAM (LTE Category 11). Figure 2 describes this evolution to LTE-A limited to 2x2 MIMO.
The evolution of LTE to LTE Advance (Figure 1 to Figure 2) introduces the principle of carrier aggregation and higher order modulation. The virtual bonding of additional LTE spectrum can be represented in this analogy as additional lanes on the freeway in order to carry more vehicle traffic or bits of information. The types of vehicle in this analogy also help to illustrate the nature of higher order modulation. A car can be thought of as 64QAM whereas the trucks are 256QAM which can carry 33% more information or cargo.
Now adding the scaling effects of spatial multiplexing described in this article, we can further extend this analogy to its theoretical limit (LTE Release 10). Figure 3 illustrates the extent of the evolved LTE Advanced network and the overall throughout yield created by 4x4 MIMO.
As we can see from this set of illustrations, pulling together key technologies such as carrier aggregation, higher order modulation and 4x4 MIMO can significantly scale up the capacity, efficiency and speed of standard LTE. By exploiting spatial multiplexing, LTE Advanced can scale with each additional carrier stream.
In real world deployments, however, LTE network operators rarely have an opportunity to deploy on large swaths of spectral holdings. Thus, Figure 3 illustrates an unlikely theoretical maximum analogy which may never be deployed. In typical LTE deployments, network carriers can usually piece together only a few frequencies at one time (i.e. the total real estate of the highway is limited to 2 or 3 lanes). Therefore, using this analogy, it is better to build up in decks of lanes using 4x4 MIMO rather than to build out in additional lanes given the limitations of spectral resource.
Another limitation of spatial multiplexing technology is, just with higher order modulation, its dependency on a high quality wireless signal. In challenging wireless communication situations, it is difficult to maintain a 4x4 MIMO on each carrier at the same time. Therefore, a more accurate illustration of a real world deployment of LTE-A with 4x4 MIMO is described in Figure 4 below.
In the above three carrier aggregation (CAT-16) scenario, two aggregated frequencies are leveraging higher order modulation as well as 4x4 MIMO and one anchor LTE frequency at 2x2 MIMO produces a grand total of 10 spatial streams. Given that each LTE frequency bandwidth is 20 MHz, this configuration would yield a LTE network that approaches the 1 Gbps performance threshold.
Where do we go from here?
In terms of technology implementation, 4x4 MIMO is clearly lagging the implementation of both carrier aggregation and higher order modulation. The primary reason for this delay is due to the engineering challenges at the device side. However, as demand for faster 4G services and higher LTE capacity continues to grow, especially driven by wireless operators with limited spectral holdings, the technology will emerge in commercial form within the next year. Chipset manufacturers such as Qualcomm have begun to roll out 4x4 MIMO capable chipsets in their Snapdragon X12 LTE modem designs. Prototype 4x4 MIMO handsets have made public demonstrations in trade shows. It is expected that network carriers, especially advanced ones in Japan, will begin deployment of 4x4 MIMO technology within the year.
In summary, by combining the three key LTE Advanced technologies of carrier aggregation, higher order modulation and 4x4 MIMO or spatial multiplexing, the industry will move inextricably closer to reaching the gigabit LTE milestone and setting up for the upcoming evolution to LTE Advanced Pro.
 Note that this 4x4 MIMO spatial multiplexing applies only to the downstream transmission from the cell tower to the user device whereas MIMO configurations made popular in the WLAN communications (WiFi) describes a symmetrical transmit and receive spatial multiplexing.