Massive MIMO Trials Pick Up Pace

In September 2017, Huawei and Bharti Airtel in India announced a partnership to deploy massive MIMO technology in one of the main urban and business centers in India. Until the beginning of 2017, the majority of massive MIMO deployment news came from SoftBank in Japan and China Mobile in China. These early systems provided by Huawei and ZTE featured 128 antenna elements (the total number of transmit and receive antennas at the base station) deployed on commercial LTE networks. In Japan, Softbank has deployed massive MIMO systems in about 100 sites across 6 cities including Tokyo to increase the speeds of its existing services. Both operators have also stated that these technologies are pivotal to future 5G deployment. However, in the first six months of 2017 more than ten trials have been reported now; the Huawei and Bharti Airtel announcement being among the most recent. 

Massive Multiple Input Multiple Output (MIMO) antenna technology is a key 5G technology that will enable operators to deliver average speeds of 100 Mbps user experienced data rate as defined by IMT-2020. While neither the 5G standards nor the spectrum bands have been finalized, some major operators globally have started massive MIMO trials or limited deployments in 2017. Among these operators are Sprint, Deutsche Telekom, China Mobile, China Telecom, China Unicom, Singtel, T-Mobile Netherlands, Vodafone Australia, Optus, and Telefónica. The spectrum bands at which these trials have taken place include 2.5 GHz, 2.6 GHz, 3.5 GHz, 1.8 GHz, and 2.3 GHz. Except for 3.5 GHz, the remaining frequencies are also allocated for LTE services in many countries. While most of these announcements are trials, T-Mobile Netherlands has activated a 64X64 (64 transmit and 64 receive antennas) massive MIMO site in Amsterdam at 2.6 GHz. Sprint too has been trialing 64X64 massive MIMO systems at 2.5GHz and plans to deploy them in 2018. While the Huawei and Bharti Airtel announcement does not specify the MIMO configuration or the spectrum band at which the trial and deployment will take place, it does indicate this step is to lay the foundation for 5G. 

MIMO antenna technology has been in use since the launch of 802.11n WiFi systems, but was first ratified for use in cellular systems in 3GPP’s Release 7 in 2008. MIMO systems essentially use multiple antennas at the transmitter and the receiver and leverage multiple propagation paths to reach the user. These multiple propagation paths or diversity in the signal path increases with multipath propagation, signal interference and noise in the channel. Multiple propagation paths are used to send the same data packets, which increases the reliability of the transmission, but if these multiple paths are used to send different data packets, it increases the bandwidth of the channel. In the latter scenario, MIMO fundamentally uses spatial multiplexing to improve spectral efficiency and hence increase the data speed.  Subsequent 3GPP releases have introduced support for higher order MIMOs in the uplink and downlink and additional key features such as beamforming (shifts in amplitude and phase to increase the strength of the signal in a specific direction), full dimensional beamforming (beamforming in the horizontal and vertical direction), user equipment support, and multi-user MIMO. In practice, however, the deployment of MIMO has been limited with 2X2 MIMO being the most common configuration today. The shift toward 4X4 MIMO has only begin with network upgrade from LTE (3GPP Release 8) to LTE Advanced (Release 10) and LTE Advanced Pro (Release 13).

Massive MIMO is an extension of the basic MIMO systems that incorporate many antennas to support very high throughput data rates, which will be critical for 5G. The exact number of antennas that would qualify a MIMO system as a massive MIMO system is debatable with some describing even 16 element antennas as massive MIMO. The crucial point is that the optimal number of antenna elements depends on the spectrum band. 5G is expected to be deployed in a range of frequencies from 600 MHz to 100 GHz. Lower frequencies can only support a limited number of antennas due to the limitation imposed by the longer wavelength and hence, very large antenna sizes. Deployments below 1 GHz are most likely to support 8 or 16 antenna elements at most. Very high frequencies above 30GHz can have hundreds of antenna elements with some research citing below 500 antennas as an upper limit. For example, infrared communications that take place above 300GHz have receiving elements that can be regarded as arrays of hundreds of receiving elements. A similar concept will apply to the mmWave bands.

Regardless of the frequency band, using multiple antennas has significance regarding costs, power consumption, and form factor, all of which increase as the number of antennas are increased, which in turn increases with the frequency of operation. Some of the other key challenges arise from the way MIMO systems use and build channel models. Before transmission of data, a pilot signal is used to gather key channel parameters and build the channel state information (CSI), which in turn helps build the channel model. This is known as the channel training time. As the number of antennas increase, the time and the amount of information that needs to be processed quickly before data transmission also increases. The CSI remains valid only for a certain amount of time known as the channel coherence time after which new CSI is required. It is also this need for accurate CSI information with minimal time lag that makes TDD spectrum more suitable for massive MIMO operation than FDD spectrum. In TDD system, uplink and downlink equivalence can be assumed using the channel reciprocity theorem since the transmission takes place over the same channel separated by time, and is therefore more accurate than in FDD systems. In FDD systems, uplink and downlink frequencies are different and channel training occurs only for the downlink and channel equivalence is assumed for uplink. Furthermore, research has shown that the training time in TDD systems increases with the number of users, whereas in FDD systems the training time depends on the number of user equipment and base stations. At high frequencies it is a challenge for the beamforming transmitter to rapidly respond to the changes in the CSI. Telecom vendors are attempting to overcome the limitations of FDD Massive MIMO by estimating the channel, for example gathering measurements from all the networks they have deployed globally and using these to predict what the CSI will be in certain conditions.

Given these set of challenges massive MIMO trials are critical for gaining a better understanding of the systems through tests and field trials. Furthermore, the implementation of these tests and trials in developing economies is crucial for adoption of 5G in the next 3 to 5 years. Even though MIMO challenges vary by spectrum band, propagation characteristics are similar for frequencies in the 1-6GHz range, within which range some of the key new 5G spectrum bands also fall. The knowledge and practical understanding gained through deployment of massive MIMOs even in LTE bands in this range can lay a solid foundation for deploying 5G networks in future.