Low-Power Design

Proposed solutions for achieving LTE-Advanced performance targets for the radio interface are defined in 3GPP TR 36.814, “Further Advancements for E-UTRA Physical Layer Aspects.” [12] A comprehensive summary of the overall LTE-Advanced proposals including radio, network, and system performance can be found in the 3GPP submissions to the first IMT-Advanced evaluation workshop. [13] The remainder of this application note will focus on the radio interface of LTE-Advanced and other Release 10 features.

The following are current solution proposals for the LTE-Advanced radio interface.

Within Release 10 there is other ongoing work that is complementary to LTE-Advanced but not considered essential for meeting the ITU requirements.

We’ll examine each of these categories from the physical layer perspective, along with some of the associated design and test challenges.

Prior to the elaboration of the Release 10 UE radio specifications in 36.101, Technical Report (TR) 36.807 [14] is being drafted. This will cover the following Release 10 features:

Like most technical reports, this document contains useful background information on how the requirements were developed which will not necessarily be evident in the final technical specifications.

The existing UE categories 1-5 for Release 8 and Release 9 are shown in Table 4. In order to accommodate LTE-Advanced capabilities, three new UE categories 6-8 have been defined. [15]

Note that category 8 exceeds the requirements of IMT-Advanced by a considerable margin.

Given the many possible combinations of layers and carrier aggregation, many configurations could be used to meet the data rates in Table 4. Tables 5 and 6 define the most probable cases for which performance requirements will be developed.

LTE-Advanced key technologies

Achieving the 4G target downlink peak data rate of 1 Gbps will require wider channel bandwidths than are currently specified in LTE Release 8. At the moment, LTE supports channel bandwidths up to 20 MHz, and it is unlikely that spectral efficiency can be improved much beyond current LTE performance targets. Therefore the only way to achieve significantly higher data rates is to increase the channel bandwidth. IMT-Advanced sets the upper limit at 100 MHz, with 40 MHz the expectation for minimum performance.

Figure 3. Contiguous aggregation of two uplink component carriers
Because most spectrum is occupied and 100 MHz of contiguous spectrum is not available to most operators, the ITU has allowed the creation of wider bandwidths through the aggregation of contiguous and non-contiguous component carriers. Thus spectrum from one band can be added to spectrum from another band in a UE that supports multiple transceivers. Figure 3 shows an example of contiguous aggregation in which two 20 MHz channels are located side by side. In this case the aggregated bandwidth covers the 40 MHz minimum requirement and could be supported with a single transceiver. However, if the channels in this example were non-contiguous—that is, not adjacent, or located in different frequency bands—then multiple transceivers in the UE would be required.

The term component carrier used in this context refers to any of the bandwidths defined in Release 8/9 LTE. To meet ITU 4G requirements, LTE-Advanced will support three component carrier aggregation scenarios: intra-band contiguous, intra-band non-contiguous, and inter-band non-contiguous aggregation. The spacing between center frequencies of contiguously aggregated component carriers will be a multiple of 300 kHz to be compatible with the 100 kHz frequency raster of Release 8/9 and at the same time preserve orthogonality of the subcarriers, which have 15 kHz spacing. Depending on the aggregation scenario, the n x 300 kHz spacing can be facilitated by inserting a low number of unused subcarriers between contiguous component carriers. In the case of contiguous aggregation, more use of the gap between component carriers could be made, but this would require defining new, slightly wider component carriers.

An LTE-Advanced UE with capabilities for receive and/or transmit carrier aggregation will be able to simultaneously receive and/or transmit on multiple component carriers. A Release 8 or 9 UE, however, can receive and transmit on a single component carrier only. Component carriers must be compatible with LTE Release 8 and 9.

In Release 10, the maximum size of a single component carrier is limited to 110 resource blocks, although for reasons of simplicity and backwards compatibility it is unlikely that anything beyond the current 100 RB will be specified. Up to 5 component carriers may be aggregated. An LTE-Advanced UE cannot be configured with more uplink component carriers than downlink component carriers, and in typical TDD deployments the number of uplink and downlink component carriers, as well as the bandwidth of each, must be the same.

For mapping at the physical layer (PHY) to medium access control (MAC) layer interface, there will be one transport block (in the absence of spatial multiplexing) and one hybrid-ARQ entity for each scheduled component carrier. (Hybrid ARQ is the control mechanism for retransmission.) Each transport block will be mapped to a single component carrier only. A UE may be scheduled over multiple component carriers simultaneously. The details of how the control signaling will be handled across the multiple carriers are still being developed.

Aggregation techniques are not new to 4G; aggregation is also used in HSPA and 1xEV-DO Release B. However, the 4G proposal to extend aggregation to 100 MHz in multiple bands raises considerable technical challenges owing to the cost and complexity that will be added to the UE. Moreover, operators will have to deal with the challenge of deciding what bands to pick for aggregation and it may be some time before consensus is reached allowing sufficient scale to drive the vendor community. 3GPP initially identified 12 likely deployment scenarios for study with the intention of identifying requirements for spurious emissions, maximum power, and other factors associated with combining different radio frequencies in a single device. However, because of the number of the scenarios and limited time, the study for Release 10 LTE-Advanced was initially limited to two scenarios, one intra-band TDD example and one inter-band FDD example. In June 2010 a third scenario was added for bands 3 and 7, as shown in Table 7. This scenario is an important combination for Europe, where re-farming of the underused 1800 MHz band currently allocated to GSM is a significant possibility.

The physical layer definition for CA is considered 80% complete and although the CA concept is simple, the details of the physical layer changes to support the signaling are complex and involve changes to the PCFICH, PHICH, PDCCH, PUCCH, UL power control, PUSCH resource allocation, and the UCI on the PUSCH. The radio performance aspects are only at 30% completion. This is significant, as Table 7 just begins to describe the possible scope of CA. To get some idea of the number of combinations requested by operators, refer to Annex A of TR 36.807. Every combination introduced into the specifications has to be assessed for aspects such as required guard bands, spurious emissions, power back off, and so forth.

One of the new challenges that CA introduces to the radio specifications is the concept of variable TX/RX frequency separation. This attribute impacts specifications for reference sensitivity and receiver blocking, among others. In Release 8 and Release 9, the TX and RX separation for each of the 19 defined FDD bands is fixed. The introduction of CA changes that, since asymmetric uplink and downlink allocations will be commonplace. The asymmetry is driven by three scenarios; different numbers of CCs in the uplink and downlink, different bandwidths of CC in the uplink and downlink, and finally a combination of different bandwidths and numbers of CCs. How to limit the allowed allocations in order to minimize the number of test scenarios is still under study.

Today’s LTE uplink is based on SC-FDMA, a powerful technology that combines many of the flexible aspects of OFDM with the low peak to average power ratio (PAPR) of a single carrier system. However, SC-FDMA requires carrier allocation across a contiguous block of spectrum and this prevents some of the scheduling flexibility inherent in pure OFDM.

LTE-Advanced enhances the uplink multiple access scheme by adopting clustered SC-FDMA, also known as discrete Fourier transform spread OFDM (DFT-S-OFDM). This scheme is similar to SC-FDMA but has the advantage that it allows noncontiguous (clustered) groups of subcarriers to be allocated for transmission by a single UE, thus enabling uplink frequency-selective scheduling and better link performance. Clustered SC-FDMA was chosen in preference to pure OFDM to avoid a significant increase in PAPR. It will help satisfy the requirement for increased uplink spectral efficiency while maintaining backward-compatibility with LTE.

Figure 4. Enhanced uplink multiple access block diagram Figure 4 shows a block diagram for the enhanced uplink multiple access (clustered SC-FDMA) process. There is only one transport block and one hybrid ARQ entity per scheduled component carrier. Each transport block is mapped to a single component carrier, and a UE may be scheduled over multiple component carriers simultaneously using carrier aggregation, as described in the previous section.

Examples of different Release 8 and Release 10 uplink configurations are given in Figure 5. The key point is that all Release 8 configurations are single carrier, which means that the PAPR is no greater than the underlying QPSK or 16QAM modulation format, whereas in Release 10 it is possible to transmit more than one carrier, which makes the PAPR higher than the Release 8 cases. Note that the multiple carriers referred to here as part of clustered SC-FDMA and simultaneous PUCCH/PUSCH are contained within one component carrier and should not be confused with the multiple component carriers of CA.

Figure 5. Comparison of Release 8 and proposed Release 10 uplink
configurations The initial specifications are likely to limit the number of SC-FDMA clusters to two, which will provide some improved spectral efficiency over single cluster when transmitting through a frequency-selective channel with more than one distinct peak.

Figure 6 shows the Release-8 LTE limits for antenna ports and spatial multiplexing layers. The downlink supports a maximum of four spatial layers of transmission (4x4, assuming four UE receivers) and the uplink a maximum of one per UE (1x2, assuming an eNB diversity receiver). In Release 8, multiple antenna transmission is not supported in order to simplify the baseline UE, although multiple user spatial multiplexing (MU-MIMO) is supported. In the case of MU-MIMO, two UEs transmit on the same frequency and time, and the eNB has to differentiate between them based on their spatial properties. With this multi-user approach to spatial multiplexing, gains in uplink capacity are available but single user peak data rates are not improved.

Figure 6. Release 8 LTE maximum number of antenna
ports and spatial layersTo improve single user peak data rates and to meet the ITU-R requirement for spectrum efficiency, LTE-Advanced specifies up to eight layers in the downlink which, with the requisite eight receivers in the UE, allows the possibility in the downlink of 8x8 spatial multiplexing. The UE will be specified to support up to four transmitters allowing the possibility of up to 4x4 transmission in the uplink when combined with four eNB receivers. See Figure 7.

The work to define the enhanced downlink is about 80% complete. There will be changes to the UE-specific demodulation reference signal (DMRS) patterns to support up to eight antennas. Channel state information reference signals (CSI-RS) and associated modifications to UE feedback in the CSI codebook design will be introduced. There also will be equivalent changes for downlink control signaling.

Figure 7. LTE-Advanced maximum number of antenna
ports and spatial layersThe specification for DMRS for Ranks 1 to 4 is given in Figure 8. DMRS support for Ranks 5 to 8 is not defined for Release 10 but is not precluded in future releases. Release 10 emphasizes dual-layer spatial multiplexing augmented by four-antenna beamsteering rather than a pure 8-layer spatial multiplexing approach, which would offer higher peak rates but require eight receive antennas in the UE.

The CSI-RS are introduced in the downlink to enable UE-specific weights to be applied to the RS for UE channel measurement purposes according to the CSI feedback. In this way the behavior of the UE-specific RS will track that of the precoded data (PDSCH), which is already optimized for each UE. The design of the CSI-RS offers other advantages over the legacy CRS in that higher reuse factors are available, which makes the introduction of inter-cell interference cancellation (ICIC) more practical. The proposed mappings of the CSI-RS for two, four, and eight antenna ports is given in Figure 9.

Figure 10 illustrates the resource block (RB) allocation for a 10 MHz FDD signal transmitted over an EPA channel as seen at the antenna of a single input UE. This particular signaling configuration was created using Agilent SystemVue along with a “beta” version of its LTE-Advanced Release 10 library.

Figure 9. Mapping of CSI reference signals (CSI configuration 0, normal
cyclic prefix) [18]The allocation shown in Figure10 is extracted from the center 12 RBs in the first two subframes of a 10 MHz FDD downlink signal. Normal cyclic prefix is employed. The first two symbols of each subframe are reserved for the PDCCH. The center of the channel has been used for Release 8 PDSCH and the outer RBs for Release 10 PDSCH. Included in the allocation are cell-specific RS along with Release 10 DMRS.

The principles for a new codebook for the 8Tx case have been agreed to, but for the 2Tx and 4Tx cases, the Release 8 codebook will be reused as it is considered good enough. However, several proposals are being considered to improve CQI/ PMI/RI accuracy for both MU-MIMO and SU-MIMO:

• Aperiodic PUSCH CQI mode 3-2 (sub-band CQI + sub-band PMI)
• Extension of Release 8 periodic PUCCH CQI mode 2-1 with sub-band PMI
• Potential enhancement on CQI for MU
• Potential enhancement on interference measurement for CQI
• UE procedure to derive PMI targeting for both MU-MIMO and SU-MIMO

Extensions of some of the Release 8 aperiodic PUSCH CQI feedback modes (1-2, 2-2, and 3-1) is proposed along with extensions of the periodic PUCCH modes 1-1 and 2-1.

Various modifications to the downlink control signaling have been agreed to including the following:

• Support of 2 orthogonal DMRS ports and 2 scrambling sequences for MU-MIMO operation
• No additional signaling to be added for the MU-MIMO case in which one RB is scheduled to more than one UE
• Additions to support the new 8Tx SU-MIMO mode dynamic switching between SU-MIMO and MU-MIMO

Equivalent work is ongoing to define multiple antenna transmission for the uplink. Note that in Release 8 and Release 9, only single antenna uplink transmission was defined, so the work in release 10 is not an enhancement
as is the case for multiple antenna downlink transmission, which was defined for four antennas in Release 8 and enhanced to 8 antennas in Release 10. A major issue is how uplink control information (UCI) will be multiplexed between two or more PUSCH. This is also an issue for carrier aggregation. Essential agreements have been reached on resource sizes for HARQ, RI, CQI, and PMI. Agreement has been reached on mapping of the PHICH on the downlink for uplink SU-MIMO, and on the cyclic shift and orthogonal cover code (OCC) definitions for the uplink DMRS. Enhancements to the sounding reference symbols (SRS) have been proposed.

Figure 10. Example of resource block allocation in LTE-Advanced The physical layer definition for multiple antenna transmission is well advanced, although the radio performance aspects for the UE and eNB are still in the early stages of discussion with completion not expected until June 2011.

Release 10 and beyond: Technologies under consideration

Coordinated multipoint (CoMP) is an advanced variant of MIMO being studied as a means of improving performance for high data rates, cell-edge throughput, and system throughput in high load and low load scenarios.

Figure 11 compares traditional MIMO downlink spatial multiplexing with coordinated multipoint. The most obvious different between the two systems is that with coordinated multipoint, the transmitters do not have to be physically co-located, although they are linked by some type of high speed data connection and can share payload data.

Figure 11. Comparison of traditional downlink MIMO and coordinated
multipoint In the downlink, coordinated multipoint enables coordinated scheduling and beamforming from two or more physically separated locations. These features do not make full use of CoMP’s potential, because the data required to transmit to the mobile needs to be present at only one of the serving cells. However, if coherent combining, also known as cooperative or network MIMO, is used, then more advanced transmission is possible.
The CoMP approach to MIMO requires high speed, symbol-level data communication between all the transmitting entities, as indicated on the right hand side of Figure 11 by a line between eNB1 and eNB2. Most likely the physical link carrying the LTE X2 interface, a mesh-based interface between the base stations, will be used for sharing the baseband data.

The coherent combining used in CoMP is somewhat like soft combining or soft handover, a technique that is widely known in CDMA systems in which the same signal is transmitted from different cells. With coherent combining, however, the data streams that are being transmitted from the base stations are not the same. These different data streams are precoded in such a way as to maximize the probability that the UE can decode the different data streams. In the uplink, the use of coordination between the base stations is less advanced, simply because when two or more UEs are transmitting from different places, there is no realistic mechanism for sharing the data between UEs for the purposes of precoding. Thus the uplink is restricted to using the simpler technique of coordinated scheduling. On the other hand, there is considerable opportunity at the eNB receivers to share the received data prior to demodulation to enable more advanced demodulation to be performed. The downside is the consequence that for a 10 MHz signal, the backhaul could be as much as 5 Gbps of low latency connections between the participating eNBs.

Simulations of coordinated multipoint have shown that when the system is not fully loaded, the CoMP process can provide substantial performance gains. However, as the load on the system increases, these gains begin to disappear. 3GPP’s recent simulation data showed initial performance improvement to be in the 5% to 15% range. This was not considered sufficient to keep coordinated multipoint as a proposal in Release 10, given the timeline for finalizing the specification. Also, recent results from the EASY-C testbed showed limited performance gains in lightly loaded networks with minimal or no interference. [19] Coordinated multipoint will be studied further for 3GPP Release 11. It remains unclear what eNB testing of CoMP might entail as it is very much a system level
performance gain and is difficult to emulate.

Another method of improving coverage in difficult conditions is the use of relaying. The main use cases for relays are to improve urban or indoor throughput, to add dead zone coverage, or to extend coverage in rural areas.

The concept of relaying is not new but the level of sophistication continues to grow. Figure 12 shows a typical scenario. A relay node (RN) is connected wirelessly to the radio access network via a donor cell. In the proposals for Release 10, the RN will connect to the donor cell’s eNB (DeNB) in one of two ways:

• In-band (in-channel), in which case the DeNB-to-RN link shares the same carrier frequency with RN-to-UE links.
• Out-band, in which case the DeNB-to-RN link does not operate in the same carrier frequency as RN-to-UE links.

Figure 12. In-channel relay and backhaul
The most basic and legacy relay method is the use of a radio repeater, which receives, amplifies and then retransmits the downlink and uplink signals to overcome areas of poor coverage. In the figure, the repeater could be located at the cell edge or in some other area of poor coverage. Radio repeaters are relatively simple devices operating purely at the RF level. Typically they receive and retransmit an entire frequency band, so they must be sited carefully. In general, repeaters can improve coverage but do not substantially increase capacity.

More advanced relays at layer 2 can decode transmissions before retransmitting them. Traffic can then be forwarded selectively to and from the UE local to the RN, thus minimizing the interference created by legacy relays that forward all traffic. Depending on the level at which the protocol stack is terminated in the RN, such types of relay may require the development of relay-specific standards. This can be largely avoided by extending the protocol stack of the RN up to Layer 3 to create a wireless router that operates in the same way that a normal eNB operates, using standard air interface protocols and performing its own resource allocation and scheduling.

The concept of the relay station can be applied in low density deployments where a lack of suitable backhaul would otherwise preclude use of a cellular network. The use of in-band or in-channel backhaul can be optimized using narrow, point-to-point connections to avoid creating unnecessary interference in the rest of the network. Multi-hop relaying is also possible, as Figure 12 shows. In this case a signal is sent from the DeNB to the first RN and then on to the next RN and finally down to the UE. The uplink signal coming back from the UE gets transmitted up through the RNs and back to the DeNB. This technique is possible to do in-channel in an OFDMA system because the channel can be split into UE and backhaul traffic. The link budget between the DeNB and the RN can be engineered to be good enough to allow the use of some of the subframes for backhaul of the relay traffic. These subframes are the ones which otherwise could have been allocated for use with multimedia broadcast in a single frequency network (MBSFN).

In Release 10 progress is being made on the RAN aspects of relaying but it is likely that the network security aspects will be delayed until Release 11. This delay may not affect RAN standardization but may impact deployment.

Release 10 intends to address the support needs of heterogeneous networks that combine low power nodes (such as picocells, femtocells, repeaters, and RNs) within a macrocell. Deployment scenarios under evaluation are detailed in TR 36.814 Annex A. [20]

As the network becomes more complex, the subject of radio resource management is growing in importance. Work is ongoing to develop more advanced methods of radio resource management including new self-optimizing network (SON) features. The Release 10 specifications also continue to develop the use of femtocells and home base stations (HeNBs) introduced in Release 9 as a means of improving network efficiencies and reducing infrastructure costs.

Today’s cellular systems are very much centrally planned, and the addition of new nodes to the network involves expensive and time-consuming work, site visits for optimization, and other deployment challenges. Some limited SON capability was introduced in Release 8 and is being further elaborated in Release 9 and Release 10.

The intent of SON is to substantially reduce the effort required to introduce new nodes and manage the network. There are implications for radio planning as well as for the operations and maintenance (O&M) interface to the base station.

• Self configuration–The one-time process of automating a specific event, such as the introduction of a new femtocell, by making use of the O&M interface and the network management module
• Self optimization–The continuous process of using environmental data, such as UE and base station measurements, to optimize the current network settings within the constraints set by the configuration process
• Self healing–The process of recovering from an exceptional event caused by unusual circumstances, such as dramatically changing interference conditions or the detection of a ping pong situation in which a UE continuously switches between macro and femto cells.

Another category of network enhancement that will figure prominently in Release 10 is the femtocell or home eNode B (HeNB).

Figure 13. Femtocell deployment in a heterogeneous 3GPP work on femtocell inclusion in UMTS was ongoing during Release 8 and was extended in Release 9 to LTE with the HeNB. In Release 9 only inbound mobility (macro to HeNB) was fully specified. Further enhancements to enable HeNB to HeNB mobility will be added in Release 10. Currently three different proposals for enabling HeNB to HeNB mobility are being studied and a decision is expected in Dec 2010. This capability is very important for enterprise deployments. Although the femtocell concept is not unique to LTE or LTE-Advanced, an opportunity exists for LTE to incorporate this technology from the start rather than retrospectively designing it into legacy systems such as UMTS and GSM. Figure 13 shows the topology of a femtocell deployment.

From a radio deployment perspective the femtocell operates over a small area within a larger cell. The radio channel could be a channel shared with a larger cell (known as co-channel deployment) or it could be a dedicated channel. The femtocell concept is fundamentally different from relaying since the femtocell connection back into the core network is provided locally by an existing DSL or cable internet connection rather than over the air back to the macrocell. Most femtocell deployments will be indoors, which helps provide isolation between the femtocell and macrocell. Also depicted in Figure 13 is a femtocell outside the macrocell coverage area. This shows how femtocells might be used to provide local cellular coverage in rural areas where DSL service exists but not
that of the preferred operator.

Although the term “femtocell” suggests that the major difference from existing systems is one of coverage area, the defining attributes of femtocells are far more numerous than coverage area alone. They include such considerations as infrastructure cost and financing; method of backhaul; network planning, deployment, quality of service, and control; mobility and data throughput performance.

The two main deployment scenarios for femtocells are in the following locations:

• In rural areas with poor or no (indoor) coverage, probably using co-channel deployment
• In dense areas to provide high data rates and capacity

In both cases operators must decide whether the femtocell will be deployed for closed subscriber group (CSG) UE or for open access. This and other practical considerations such as pricing can be considered commercial issues, although in the co-channel CSG case, the probability that areas of dense femtocell deployment will block macrocells becomes an issue.

The potential gains from femtocells are substantial, but they present many challenges. Solutions are needed for many of the following, some of which are being addressed in Release 10:

• Cognitive methods to reduce interference to the macro network
• Radio resource management requirements
• Methods of addressing security concerns associated with users building their own cellular networks
• Verification of geographic location and roaming aspects
• Business models for open- versus closed-access operation
• Support of more than one network per femtocell
• Ownership of the backhaul and the issue of net neutrality
• Optimized and balanced interworking between macrocells and femtocells to minimize unnecessary handovers
• Methods of resolving bottlenecks on fixed broadband backhaul connection, especially on the uplink for services requiring symmetric bandwidths, prioritization, and congestion management
• QoS control for real-time services (such as voice) and applications requiring guaranteed bit rates
• Access control providing closed subscriber group local and roaming access
• Capability for self-configuration, self-organization, self-optimization, and selfhealing (including fault management and failure recovery)
• Security, backhaul protection, device and user authentication

Table 8. Comparison of macrocell/microcell and femtocell/hotspot use models In spite of these issues, studies have shown that increases in average data rates and capacity of some 100x are possible with femtocells over what can be achieved from the macro network. On the other hand, femtocells do not provide the mobility of macrocellular systems, and differences exist in the use models of these systems, as shown in Table 8. For these reasons, femtocell and hotspot deployments should be considered complimentary to rather than competitive with macrocells and microcells.

Customer premises equipment in the context of the 3GPP specifications refers to a UE in a fixed location. Two main deployment scenarios are given in TR 36.807, as shown in Figure 14.

The main advantage of the CPE is that it can be optimally located using a higher performance antenna, and it is defined with a higher output power of up to 27 dBm compared with 23 dBm for a standard UE. Customer premises equipment is also less likely to be battery powered, which gives added design freedom to optimize
radio performance. The indoor scenario will likely involve an omni-directional antenna whereas the outdoor scenario will likely be deployed using some form of directional antenna.

The combination of antenna positioning, output power, fixed location, and less concern about power consumption dramatically changes the performance that would be possible using a typical mobile UE. This extra radio performance is particularly useful where LTE might be used to provide high performance broadband services; for example, in rural areas. Such deployment is seen as an attractive use of the “digital dividend” spectrum freed up by the switchover from analog to digital television.

References

Insert your comment

LTE-Advanced and Other Release 10 Solution Proposals

LTE-Advanced key technologies

Release 10 and beyond: Technologies under consideration

References