Cooperative Communication Based on Spectrum Aggregation in LTE Further Evolution

This work was supported by the National Science and Technology Major Project of the Ministry of Science and Technology of China under Grant No. 2009ZX03003.

    The advantage of using large-bandwidth wireless transmission is its high data rate, which can support multimedia services. The indirect advantage is that it lowers the receiver’s power consumption by shortening data transmission time. The combination of large-bandwidth wireless communication and multimedia terminals can also change traditional service modes. For example, media in traditional Video on Demand (VoD) and video broadcast, except those for live broadcast, can all be downloaded instantly through large-bandwidth transmission to the local device and then be played back. This mode increases the flexibility of watch time, place, and content, and lowers the power consumption of the terminal receiver and display. In addition, this service mode can lower the requirements of the seamless coverage on the
large-bandwidth wireless network, thus lowering the cost of network construction.

    With so many advantages, large-bandwidth wireless transmission has become the mainstream of mobile communications. The transmission bandwidth of mobile communication systems have increased continuously from the 5 MHz (initial bandwidth) of the Universal Mobile Telecommunications System (UMTS) to 20 MHz of the Long Term Evolution (LTE) system, and to
100 MHz of Long Term Evolution Advanced (LTE-A).

    There are two basic solutions for the large-bandwidth transmission in mobile communication systems: the first is to design a large-bandwidth system; the second is to construct a system with a larger transmission bandwidth through collaboration between different systems. The two solutions coexist and have an effect on each other in the evolution of mobile communications. The first was adopted in the discussion of the 3GPP LTE-A standard, and is used in designing new systems. The second is a cost-effective scheme for carriers to evolve their existing networks.

    The two solutions mentioned above, which obtain larger transmission bandwidth through carrier aggregation, are cooperative communication based on spectrum aggregation. Reference [1] makes many biological analyses of the cooperative communication, but lacks analysis of bionomic cooperation. It also lacks analysis of cooperative communication in terms of spectrum aggregation. This article analyzes spectrum aggregation in terms of cooperative communication, which aids understanding of a broadband system design and the evolution of carriers’ existing networks. It also seeks a solution to the problems in the evolution of existing networks.
This article mainly discusses  cooperative communication based on spectrum aggregation among different systems, especially cooperative communication among different systems through heterogeneous spectrum aggregation. This provides a solution to problems that existing multi-carrier binding technologies cannot solved.

1 Spectrum Aggregation and Cooperative Communications

1.1 Current Development of Spectrum Aggregation
Before research on the standard for the 4G mobile communication system LTE-A has begun, the research on carrier aggregation had been started or finished on the protocol layer in 2G and 3G mobile communication systems, as shown in Figure 1. Representative carrier aggregation technical specifications include the Time Division Synchronous Code Division Multiple Access
(TD-SCDMA) system, and the Data Optimized (DO) Multi-Carrier Multilink Extension (DMMX) and High-Speed Data Packet Access (HSDPA) Multi-carrier Multilink Extension (HMMX) platforms launched by Qualcomm. These platforms have been developed to support the long-term evolution of Evolution Data Optimized (EV-DO) and HSDPA technologies.

    The 2G, 3G and 4G spectrum aggregation solutions given in Figure 1 are all achieved through carrier aggregation. In spectrum aggregation used in the 2G mobile communication system shown in Figure 1 (b), Qualcomm DMMX and HMMX support multi-carrier and multi-link transmission, which allows multiple wireless transmission protocols on many frequency bands. For example, the Media Forward Link Only (MediaFLO) link based on  Orthogonal Frequency Division Multiplexing (OFDM) at 700 MHz, which is used for video service, and the EV-DO backward link based on the CDMA at the cellular band, form a platform that provides support for inter-system (or cross-protocol) spectrum aggregation.

    For the spectrum aggregation used in 2G and 3G mobile communication systems, only Qualcomm DMMX and HMMX support cross-band, cross-protocol carrier aggregation. Other systems, such as GSM, TD-SCDMA and the multi-carrier HSPA for UMTS, all use intra-system continuous carrier aggregation. These systems have only one objective, to expand transmission bandwidth. However, the evolution of LTE carrier aggregation is included in 4G LTE-A. For LTE-A, although the range of carrier aggregation is expanded from that of 3G continuous carriers to that of
non-continuous carriers, it is still limited to the intra-system carrier aggregation. Currently, LTE-A has not considered the support of inter-system carrier aggregation. The inter-system spectrum aggregation in 4G spectrum aggregation shown in Figure 1 (c) indicates its technical feasibility.

1.2 Development Trend of Spectrum Aggregation
The following basic problems of LTE-A are related to the cooperative communication based on spectrum aggregation:

    (1) How to obtain large-bandwidth spectrum
    Is large transmission bandwidth achieved through carrier aggregation within the same LTE-A FDD or LTE-A TDD system, or through cooperation between the two systems by means of carrier aggregation?

    (2) How to use spectrum effectively
    Services requiring large bandwidth are data services with the distinctive characteristic of upstream and downstream asymmetry, varying depending on time and place. How does multi-carrier aggregation in LTE-A FDD adapt to such asymmetry? This also involves whether to use inter-system carrier aggregation and whether to aggregate the spectrum of the LTE-A FDD and LTE-A TDD systems to support asymmetric services together.

    The method of multi-carrier binding spectrum aggregation on the same type of continuous spectrum is not a solution for International Mobile Telecommunications Advanced (IMT-A). It differs to ensure the efficiency of spectrum usage in the cases of asymmetric services through the aggregation of FDD spectrum only. It is also different to solve the high TDD feedback delay and scheduling delay problems (restricted by the wireless frame structure) by using continuous carrier aggregation on the TDD spectrum. These are the factors limiting the further improvement of spectrum efficiency.

    In addition, insufficient low-end spectrum makes it very difficult to provide each carrier with a separate spectrum of up to 100 MHz (even if there is sufficient bandwidth, it cannot be fully used). Therefore, high-end spectrum is required to dynamically supplement the low-end spectrum used for macro coverage, and to expand the applicable scenarios of high-end spectrum. This requires more flexible cooperative communication based on spectrum aggregation. Simple carrier binding cannot solve the problems.

    Different methods of spectrum aggregation can solve different problems. Flexible spectrum aggregation can expand transmission bandwidth, introduce new services, improve the spectrum use of air interfaces and expand the application scenarios of high-end spectrum. In all the methods of spectrum aggregation, that between different systems, between
non-continuous spectrum and the aggregation of high-end and low-end spectrum can often solve problems that cannot be solved in cases of traditional spectrum aggregation.

1.3 Cooperative Communications Based on Spectrum Aggregation
Cooperative communication provides functions that a single communication function entity does not have, through the cooperation of a group of entities. In cooperative communication based on spectrum aggregation, communication function entities are functional or physical entities that can transmit and/or receive radio signals on a single carrier. If the cooperative function entities come from different systems, it is inter-system communication.

    To simplify the variety of networks and lower network construction costs in the evolution of existing wireless access networks, base-stations, and the transmission parts in different wireless access networks are converged gradually. However, existing terminals using different air interfaces are difficult to converge, resulting in a variety of air interfaces in the existing wireless access networks in a long term. Cooperative communication based on spectrum aggregation can provide complementary advantages among different systems in cases where there are a variety of air interfaces.

    Moreover, cooperative communication based on spectrum aggregation can be centralized management/control-based, distributed management/control-based and Ad hoc management/control-based in terms of implementation. However, all modes of cooperative communication are based on radio environment information. Therefore, radio environment recognition technology is closely related to the cooperative communication based on spectrum aggregation. There will be more requirements of radio environment information when the inter-system cooperation based on spectrum aggregation becomes closer and the ad-hoc level becomes higher.

2 Cooperative Communication Based on Spectrum Aggregation and Asymmetric Service Support

2.1 Characteristics of Asymmetric Services
Reference [1] analyzes the asymmetry of 3G services in terms of service asymmetry, transmission flow asymmetry and spectrum asymmetry. It also analyzes the dynamic characteristics of asymmetric services in terms of user, cell and system. These analyses indicate that the asymmetry of link-level services is highly dynamic and changes greatly depending on time and space. The asymmetry of cell-level services is moderately dynamic and changes moderately depending on time and space. The asymmetry of system-level services is not dynamic and changes slowly depending on time and space.

    The asymmetry of various services described in Reference [2] indicates that a service on the mobile communication system is a comprehensive reflection of these symmetrical and asymmetric services. It has both smooth, symmetrical service flow components (corresponding to direct current components) and asymmetric, sporadic and high Peak-to-Average Ratio (PAR) components (corresponding to alternating current components) Resources are distributed in the unit of one cell or multiple adjacent cells for mobile communication services, so changes of service asymmetry within a cell are the most important reference for spectrum usage. That is, the mobile communication systems should obey the following rules when considering uplink and downlink spectrum resource distribution for the cell: uplink and downlink resources in cell unit should be distributed and should adapt to moderate asymmetry changes of uplink and downlink services.

2.2 Performance Difference Between TDD and FDD in Asymmetric Services
According to the analyses in Reference [3], both the uplink and downlink service asymmetry and PAR characteristics are different in commercial areas, residential areas, and business areas. The TDD system can dynamically adapt to service asymmetry and outburst. Therefore, in terms of a realistically attainable system capacity (where the bandwidth of the TDD system and sum of the uplink and downlink bandwidths of the FEE system are the same), the capacity of the TDD system is 69% higher than that of the FDD system due to the difference of service asymmetry. Only when the ratio of the uplink/downlink service data rate is 33.3% and 42.5%, does the FDD system have the same capacity as the TDD system. In other service cases, the capacity of the FDD system is always lower than that of the TDD system.

    Regardless of the difference between TDD and FDD in other aspects, FDD is more suitable for services with a symmetrical uplink and downlink and low PAR in relation to the type of service only, whereas TDD is suitable for service time-dependent changes, where the uplink and downlink are asymmetric.

    Reference [1] indicates the unpredictability of the service mode and the changes in terms of time and space; it is not feasible to allocate fixed asymmetric spectrum for the FDD system in spectrum distribution. Spectrum planning for FDD in LTE-A should also be the same as the traditional allocation for FDD, which means using the uplink and downlink symmetry mode. Adaptation to asymmetric services is achieved through combination with TDD or through the dynamic sharing of spectrum with other systems.

2.3 Improving Asymmetric Services Supporting FDD through  Cooperative Communication Based  on Spectrum Aggregation
The support of the FDD system for asymmetric services has been discussed in NGMN P-BAG and 3GPP LTE-A, and can be summarized in the following three solutions:

    (1) Planning asymmetric spectrum. Breaking through the traditional mode of planning symmetrical bands for uplink and downlink to improve support of the FDD system for downlink services in the spectrum planning stage, allocating a downlink band with a bandwidth higher than that of the uplink in the FDD system.

    (2) TDD spectrum used for the downlink transmission in the FDD system. To improve the downlink service capability in the FDD system, the spectrum of TDD can be used for deploying the downlink channels in the FDD system, thus increasing downlink transmission bandwidth.

    (3) Cooperative communication based on spectrum aggregation between FDD and TDD systems. A characteristic of this solution is that the TDD spectrum is used with air interfaces for the TDD system; FDD air interfaces are used in the FDD system, and based on that, carriers are aggregated between TDD air interfaces and FDD air interfaces.

    Problems related to asymmetric spectrum planning are:
    (1) How much higher should the downlink bandwidth be than the uplink bandwidth to meet the asymmetry requirements for services in the FDD system?

    (2) Asymmetric services run in a unit of cells and change in relation to  time and place. So how does the pre-planned asymmetric spectrum for the uplink and downlink adapt to such changes?

    A research report from the Information Society Technologies (IST) of the European Union pointed out that there is no solution for predicting the asymmetry of future services. Therefore, there is no theoretical solution to the problem of FDD system adaptability to asymmetric services, and so a spectrum planning solution cannot be realized.

    TDD spectrum used for the downlink transmission in the FDD system has the same problem as asymmetric spectrum planning. Given that service asymmetry cannot be estimated at a specific place and time, how much TDD spectrum should be used reasonably for transmitting FDD channels? Configuring FDD devices on TDD spectrum virtually means adding a band of spectrum to the FDD system. Reference [1] discusses such a scenario, and concludes that it is not feasible.
Figure 2 illustrates the implementation of cooperative communication between the FDD and TDD systems, in which TDD air interfaces are used on the TDD spectrum and FDD air interfaces are used on the FDD spectrum. Based on the spectrum deployment, the proportion of uplink and downlink time slots in radio frames in the TDD system can be adjusted flexibly according to the proportion of uplink and downlink service asymmetry in specific cells at specific times. Communication with specific terminals can be achieved through proportional adjustment and parallel transmission via TDD and FDD air interfaces.

    Figure 2 illustrates cooperative communication between FDD and TDD systems based on spectrum aggregation, which has the following characteristics:

• FDD devices are used on TDD spectrum
• TDD devices are used on FDD spectrum

    This solution does not involve spectrum planning or require TDD and FDD spectrum refarming, which has the following effects:

• Flexible support for sporadic and asymmetric services
• Flexibility to changes in time- and space-dependent asymmetry
• High spectrum utilization or high system capacity

    Cooperative communication based on spectrum aggregation between FDD and TDD systems improves support of the FDD system for asymmetric services using the flexible uplink and downlink service capabilities of the TDD system. It avoids the problem of predicting asymmetric services and is a robust solution with high adaptability. This solution fully utilizes the advantages of both the TDD and FDD systems and combines these advantages to support different services. The two systems cooperate closely and complement each other. In addition, this solution promotes the coexistence of TDD and FDD and the growth of TDD industrial chain, as it relates to industrial development.

3 Cooperative Communication Based on Spectrum Aggregation and Guard Band Utilization

3.1 Analysis of Guard Band Scenarios Between TDD and FDD
To save network construction costs, carriers need to share network resources. This includes sharing sites, spectrum and even antennas between different systems. With this trend, carriers require a solution for sharing the sites of base stations between TDD and the existing FDD system. This requires analyzing and solving the interference problems between TDD and FDD system in the site-/antenna-sharing mode.

    In the traditional TDD system, the uplink and downlink use the same frequency band. To prevent transmission and reception of base stations and terminals in the TDD system from being interfered with by the base stations and terminals in systems at adjacent frequency bands, or to ensure the interference is in an acceptable range, a guard band should be reserved between the TDD and FDD systems. Where a TDD network and  FDD network are constructed on different sites, such a guard band is approximately 3 MHz. Where the networks share the same site or antennas, the guard band should be higher than 10 MHz. Therefore, it is necessary to analyze the utilization of the large guard band.

    Logically, no matter which band the TDD is located in or which standard is used on the TDD band, the TDD band and its adjacent or related bands are arranged in seven forms as shown in Figure 3. Corresponding to each form of TDD/FDD spectrum arrangement, the TDD system uses the interference suppression measures, as shown on the right side of Figure 3, in spectrum usage. The seven forms of TDD/FDD spectrum arrangement illustrated in Figure 3 cover all the potential possibilities between TDD and FDD spectrum (including non-mobile communication spectrum). These forms of arrangement allow us to make an overall evaluation for the adaptation of each mode of TDD duplex.

3.2 Utilizing Guard Bands Between TDD and FDD Systems Through Cooperative Communications Based on Spectrum Aggregation
Figure 4 presents a form of arrangement of TDD and FDD, in which Band 1 is the FDD downlink band; Band 2 is the band for both uplink and downlink of the TDD system; Band 3 is the guard band between the downlink bands of the TDD and FDD systems; Band 4 is the FDD uplink band; Band 5 is the guard band between the uplink bands of the TDD and FDD systems[4].

    Reference [5] provides the following method of utilizing the guard bands: Associate Band 3 and Band 5 into a pair of HD-FDD links. Specifically, the HD-FDD system working within Band 3 provides the first HD-FDD channel. The transmission of the first HD-FDD channel is synchronized with the TDD uplink or downlink transmission. The second FDD channel is configured on the fourth band.

    Besides the utilization of guard bands through half-duplex FDD provided in Reference [5], guard bands can also be used more flexibly in terms of the cooperative communication based on spectrum aggregation. Specific solutions are as follows:

    (1) Expanding Band Through Intra-System Cooperation
    The spectrum between Band 3 and Band 2 in Figure 4 is aggregated to expand the TDD downlink transmission bandwidth, or the spectrum between Band 5 and Band 2 is aggregated to expand the TDD uplink transmission bandwidth.

    (2) Expanding Band Through Inter-System Cooperation
    As illustrated in Figure 4, spectrum is aggregated between Bands 3 and 2 of the TDD system and Band 1 of the FDD system to expand the downlink transmission bandwidth. This scheme allows the utilization of TDD guard bands and in addition, improves the support of FDD for asymmetric services. Another scheme involves aggregating spectrum between Bands 5 and 2 of the TDD system and Band 4 of the FDD system. This scheme allows utilization of TDD guard bands and in addition, improves support of FDD for asymmetric services.

4 Conclusions
Inter-system cooperative communication based on spectrum aggregation allows expansion of the transmission bandwidth for air interfaces; in addition, it can solve problems that cannot be solved by a single system. This paper discusses ways in which the FDD system can provide support for asymmetric services through inter-system cooperative communication based on spectrum aggregation. This paper also discusses solutions for guard band utilization problems caused by networks constructed on shared sites.

    In network evolution, LTE and its evolved systems will coexist with UMTS and GSM for a long time. To share network resources and lower the cost of network construction, cooperative communication is required on air interfaces between different systems. The most direct and effective solution for the cooperation is inter-system cooperative communication based on carrier aggregation, which allows multi-mode and multi-band parallel transmission.

    In current discussions on the 3GPP LTE-A standards, it is still accepted that the purpose of spectrum aggregation is to establish a single system with a 100 MHz transmission bandwidth. The spectrum aggregation discussed is the aggregation within a single system. Reconfigurable Radio Systems (RRS), a technical committee of ETSI, is a standardization organization focused on researching inter-system cooperative communication. It aims for scientific cooperation of the existing or future wireless communication systems and thus achieving cooperative communication bionomically. As the requirement of resource sharing necessary in the evolution of carriers’ existing networks increases further, inter-system cooperative communication based on spectrum aggregation will become a more prominent topic of discussion among relevant standardization organizations.

References
[1] FITZEK F H P, KATZ M D. 无线网络中的合作原理与应用 [M]. 程卫军, 等译. 北京: 机械工业出版社, 2009.FITZEK F H P, KATZ M D. Cooperation in wireless networks: principles and applications [M]. translated by Cheng Weijun, et al. Beijing: China Machine Press, 2009.
[2] MALKOWSKI M. Spectrum efficient multicast and asymmetric services in UMTS [R]. IST-2001-35125 (OverDRiVE) D08. 2003.
[3] TDD versus FDD access schemes [R].IST-1999-11571 EMBRACE D8. 2001.
[4] Next generation mobile networks spectrum requirements update [R]. A White Paper Update by the NGMN. 2009.
[5] GORMLEY E F. Utilizing guard band between FDD and TDD wireless systems [P]. US20070286156. 2007.

[Abstract] The implementation of wide band radio transmission may occur in two parallel ways: designing a new system with larger bandwidth, or constructing a cooperative system based on existing systems. The former is a spectrum aggregation scheme based on
multi-carrier transmission within one system, while the latter fulfills spectrum aggregation based on multi-carrier transmission among systems. On the one hand, the advantage of the multi-carrier transmission scheme within a system and among systems is that the existing Radio Frequency (RF) techniques and elements can be sufficiently utilized, so system costs can be decreased. On the other hand, spectrum aggregation among systems, especially among heterogeneous air-interfaces, can not only expand the basic bandwidth functions, but also fulfill efficient spectrum usage for Frequency Division Duplex (FDD) system under the unbalanced Uplink/Downlink (UL/DL) service scenario. It can also efficiently guard band usage between FDD and Time Division Duplex (TDD) systems. Commercially, this has significant implications for operators.