Carrier Aggregation, 256 QAM and now four downlink receive antennas – providing up to four parallel MIMO streams – are novel features in LTE, introduced by 3GPP to meet the demand of ever increasing data rates.
In this post we will discuss important aspects of migrating the User Equipment from two to four receive antennas in the downlink. We will explain the physical aspects, shed some light on certain important implementation issues, and identify the benefits, both from an operator’s and an end user’s perspective.
Tempting as it may be, to aggregate the performance from each of the above features, they should instead be considered as complementing each other in achieving overall higher capacity. Of course, some combinations will also appear. For example, four downlink receive antennas (4 RX) in combination with 256 QAM may turn out to be quite beneficial in high signal-noise ratio (SNR) indoor scenarios.
MIMO allows multiple data streams to be transmitted in parallel over the same physical time-frequency resource, similar to allowing vehicles in multiple lanes moving in parallel from one point to another, see Figure 1. The number of parallel streams are, however, limited to the lesser of the number of transmit antennas and the number of receive antennas. Quite often, in practice, the number of streams is even fewer than that, the actual value depending on several factors such as channel environment, as well as the antenna design of both the base station (BS) and the user equipment (UE). By adding more receive antennas, not only does the theoretical maximum number of streams increase, but also the likelihood of being able to use more streams in practice.
Figure 1: Example of four parallel, bi-directional streams. In this case, and often the case in cellular systems, the downlink is more utilized than the uplink..
Physical aspects of more receive antennas
The physical implications of increasing the number of receive antennas are twofold. First, the UE will be able to receive more signal energy, hence allowing it to receive a signal from a farther distance. Second, by using more antennas, the UE may benefit from the increased spatial diversity in order to increase sensitivity in certain (signal) directions and cancel out other (interference) directions. This may be translated into some different practical benefits that we will now go into more detail about.
Shannon’s theorem on the capacity of a channel states that the capacity, C, is bounded by the bandwidth, B, and the per stream SNR, S/N, as
Based on this relation, it is possible to identify two regions of the channel capacity: the power limited region, at low SNRs, in which predominantly the signal power limits capacity; and the bandwidth limited region, at high SNRs, in which predominantly the available bandwidth limits capacity. 4 RX will contribute in different ways for both of these regions.
In the power limited region, S/N<1, an increased number of receive antennas will correspondingly increase the received signal energy by using beam forming or spatial diversity, thereby increasing capacity. In the bandwidth limited region, S/N>1, the increased number of receive antennas will allow for more parallel streams to be received. It should be noted though, that the signal energy in one stream will appear as noise when linearly demodulating the other streams, which is why, in practice, the channel will limit the number of streams. A non-linear successive interference cancellation receiver will handle multiple streams demodulation better, since it is able to cancel the interference from the other streams.
4 RX network effects
From a consumer’s perspective, the increased peak data rates implied by 4 RX may be the discriminating feature when deciding on a new phone. However, there is much more to 4 RX than that. From the points of view of end users and operators, the increased number of receive antennas is welcome for several [other] reasons as well.
4 RX will allow for both diversity and beamforming gains that will contribute to an increased coverage. Downlink cell coverage is limited by the UE being able to receive synchronization and control signals. By doubling the number of receive antennas, this limit may be moved a couple of dBs lower which, in turn, implies a substantial increase in cell area.
In earlier releases of LTE, where, in general, only two receive antennas were used in the BS, the cell range is usually limited by the uplink. Therefore, in order to increase cell range for earlier LTE releases, a corresponding increase in BS receive antennas is necessary. Increasing the number of antennas in the BS is necessary anyway, in order to take advantage of the improved MIMO capabilities of a 4 RX UE.
Instead of assuming a wider cell range, unchanged cell coverage could be maintained by reduced transmission power in the BS. With a 4 RX UE, a reduced BS output power is sufficient to maintain the same cell coverage. The size of the reduction depends greatly on the environment. From an operator’s OPEX perspective, a reduced output power is expected to be of particular interest in existing networks. Of course, the full benefit of this will not apply until all UEs are 4 RX UEs.
One incident that rarely occurs, but is so evident and irking when it does, is a dropped link. Most often, a dropped link arises from poor radio resource management (RRM), e.g., during cell handover. More receive antennas imply that the margins at handover between two cells will increase such that the likelihood of a dropped link is reduced.
Another benefit from more RX antennas, which is also related to the dropped link, is the ability for functional parallelism, e.g., simultaneous data reception and RRM measurement. This ability could substantially reduce the time for mobility or location measurements since it eliminates the need for performing such measurements during the defined 6 ms measurement gaps for LTE. For hand-over to a GSM network, this implies cell detection in 60 ms, a procedure that may take up to several seconds using gaps. Since gaps also affect control signaling, such as HARQ or UL scheduling grants, also these are likely to benefit from the abolishment of measurement gaps.
In general, control signaling, such as PDCCH, PHICH and PCFICH, will benefit from the added receive antennas. Although they only use one MIMO stream, spatial diversity gains from the added antennas will allow for expanded cell coverage and more reliable operations at cell edges. This improves functions like downlink scheduling decisions, uplink scheduling grants and HARQ, which in turn provide secondary effects on data performance.
Link and system capacity
Peak data rate is a parameter that has been driving, and most likely will continue to drive, UE development. Although most often presented as laboratory performance figures, consumers still use it as a differentiating factor when replacing an old UE with a new. However, hand in hand with theoretical figures go the 5th and 50th percentile figures, which are often used in system performance assessment. Outdoor measurements show cell edge gains of 113% moving from 2×2 to 4×4. Indoor system simulations show that relative throughput may increase fourfold in a 4×4 (4 TX and 4 RX antennas) setup compared to a 2×2 setup, see Figure 2.
Figure 2: Relative throughput gains at 5tha dn 50th percentile compared to a 2×2 configuration.
The majority of the gains are found in the lower percentiles, i.e., where reception is the poorest. This is where the benefit from an increase is the highest in order for an operator to be able to provide consistent app coverage throughout the cell area. Consequently, the doubled throughput that is made possible with 4 RX will have an impact on operators as well as on chipset and UE vendors.
In order to achieve higher peak rates, it is necessary that also the network supports the increased number of parallel streams. The four stream MIMO transmission capability will be particularly important for operators lacking band combinations for CA, since it gives them an opportunity to increase data rates using a single band.
Quite likely, the main benefit from increasing the number of receive antennas will be the enhanced ability to simultaneously receive parallel data streams and suppress interference from neighboring cells by using an interference rejection combining (IRC) receiver, see
Figure 3: Illustration of the ability of interference mitigation in a four antenna IRC receiver by receiving two streams from two directions while simultaneously rejecting another two directions from a neighboring cell.
From the user perspective, this will be perceived as generally higher data rates, since a decreased total interference level will allow not only more parallel streams, but also higher modulation orders and lower coding schemes for those streams. For the operator, it means tighter deployment possibilities, without having to consider all neighboring cell interference.
Design aspects of 4 RX UEs
The success of 4 RX UEs will very much depend on UE vendors’ ability to optimize performance of 4 RX UEs. One challenge is to fit more antennas into the already crammed space that is the interior of a mobile phone. Another challenge is how handle the increased power consumption that 4 RX, unless properly managed, most likely will cause.
Antenna correlation, beamforming and spatial diversity
The behavior of a 4 RX UE depends to a great extent on the antenna implementation and more specifically, the correlation between the antenna elements in a multi-antenna receiver. Generally, the more widely spaced (wavelength normalized) the elements, the greater the spatial receive diversity, i.e., the UE’s ability to receive multiple, uncorrelated streams of data. In the case of few MIMO streams this translates into fading diversity such that uncorrelated antennas are less likely to suffer from fades simultaneously. When there are more MIMO streams, these can be received in parallel, allowing for higher data rates.
One way of achieving low correlation among the receive antennas is using a cross-polarized antenna configuration. In this case, the antennas are designed to be sensitive to different polarizations. The result is lower correlation with more closely spaced antennas. Antenna design and positioning will likely be a vital design parameter, since the already very confined UE space will require low antenna correlations in order to fully utilize in practice the increased number of parallel streams that a 4 RX UE will support in theory.
Although beneficial during data reception, the use of four receive antennas comes at a cost in terms of increased power consumption. An increased number of receive antennas, generally results in higher data rates, implying shorter reception time. This, in turn, all other things being equal, translates into lower power consumption. However, in non-active mode, power consumption cost may persist without the reception benefits. For this reason, UEs are likely to be equipped with a decision algorithm regarding whether to use all four antennas or to reduce the active set, depending on the present situation, in order to preserve power.
LTE introduced the multi-antenna user equipment (UE) in cellular systems already from its inception. Although not explicitly stated, the requirements defined for Release 8 tests are impossible to pass without two receive (RX) antennas.
3GPP’s RAN Working Group (WG) 4 is responsible for defining the requirements for transmission and reception parameters, and for channel demodulation performance. Although the functionality of up to eight MIMO streams has been specified in RAN WG 1 since Release 10, it is not until the ongoing Release 13 that the work on the requirements specification was initiated for up to 4 MIMO streams.
3GPP is presently standardizing the performance requirements of 4 RX, and is expected to be finished by June 2016.
By adding more receive antennas in the UE, significant gains can be achieved in several key areas. The main drawback, power consumption, will most likely be handled on an implementation basis such that UEs will only utilize the larger number of receive antennas when it is worthwhile. The benefits, in terms of increased system capacity, maximum data rates and network stability are likely to significantly affect the perceived user experience positively.
Magnus Åström, PhD
Senior Researcher, Ericsson Research.
Fredrik Nordström, PhD
Senior Researcher, Ericsson Research.