Is LTE more of a revolution than an evolution. By embracing the technological fruits of decades of academic research and practical experience, LTE represents a momentous step forward in the possibilities for communication on the move.Wider bandwidths, multicarrier modulation, multiple antennas and packet scheduling, to name but a few of the technologies finding exploitation in LTE, together provide the basis and opportunity for exciting new services and applications. Yet, however great the advance made by LTE, further steps of enhancement must surely follow.
Such ongoing enhancement is fuelled on the one hand by the need of the mobile telecommunications industry for the ability to offer new services, and on the other hand by the unabating zeal of technologists and researchers to improve on what has gone before. The provision of new services in the sphere of mobile communications is often driven by the possibilities in the fixed networks. An ever-present desire exists to achieve wirelessly the same as can be done via copper cable or optical fibre. This can in principle be done in two ways – either by using more radio spectrum, or by using the available spectrum more efficiently. The latter is heavily constrained by the laws of physics, and many aspects of LTE come close to achieving the fundamental limits of what is theoretically possible within a given bandwidth. Against this background, the International Telecommunication Union (ITU) has taken steps to ensure that more radio spectrum will be available, globally if possible, for systems beyond LTE.
One clear challenge is in adapting to the characteristics of the available frequency bands. While data rates of 1 Gbps might ideally be achieved using contiguous bandwidths of 100 MHz or more, it is clear that competition for spectrum, and the fragmentation of the available spectrum, makes it unrealistic to expect such large contiguous bandwidths in most cases. The systems of the future will therefore have to rely on non-contiguous spectrum allocation and bandwidth aggregation methods if they are to achieve the target peak data rates. This is potentially even more challenging for implementation and system design than the variable bandwidth characteristics of LTE
LTE-Advanced will not, however, be able to start from a clean sheet of paper. Investments already made in LTE have to be taken into account, and the next enhancements will be required to be backward-compatible with the first version of LTE. This means that the first LTE terminals should be able to operate in LTE-Advanced cells just as if they were LTE cells. This is analogous to the enhancement of the first version of UMTS by means of High Speed Packet Access (HSPA), whereby network operators are able to continue serving existing customers while their network equipment is progressively upgraded to enable users with newer equipment to benefit from the very latest technology. Many aspects of LTE already facilitate such backward-compatible enhancement. For example, the basic multiple access schemes and many aspects of the physical layer and protocol design (such as modulation schemes and channel structures) are easily extensible to higher data rates and wider bandwidths. Most of the characteristics of the LTE radio interface will therefore be retained, so that from operational and performance perspectives the effort to upgrade to the advanced version is kept reasonable.
It is important to note that although the use of larger bandwidths (whether contiguous or aggregated from separate carriers) may allow for higher peak data rates, it does not result in improved spectral efficiency (or reduced cost). It is in this area that further advances in theoretical understanding will continue to provide the key to enhancing the value of a given section of spectrum. Some advances can be made by extension of schemes already initiated in LTE, especially in the area of Multiple-Input Multiple-Output (MIMO) transmission. The first version of LTE assumes the presence of up to four antennas at the eNodeB and either two or four at the User Equipment (UE) (depending on the UE category). Higher order spatial multiplexing, with larger numbers of parallel data flows may be considered. One strategy consists of increasing the number of antenna elements (for example, up to eight transmit and four receive antennas) in order to increase the multiplexing gain by a factor of at least two. Such methods will be dependent on achieving sufficient decorrelation between the antennas – a potentially significant challenge on a small mobile terminal, but more feasible on a laptop computer. Another strategy would be to support more advanced MIMO techniques, perhaps making better use of knowledge of the radio channel state, in order to approach more closely the theoretical performance limits than is possible in the first version of LTE.
Other steps forward in spectral efficiency may be derived from as-yet untapped technologies. Two prime examples of these are coordinated MIMO schemes and the use of relay nodes. Coordinated MIMO schemes involve multiple cells cooperating to improve the overall spectral efficiency. Multiple eNodeBs may schedule transmissions simultaneously to one or more UEs with various aims depending on the chosen strategy. For example, coordinated scheduling may be used to reduce interference or to achieve spatial multiplexing gain by benefiting from macro-diversity resulting from the low correlation between geographically diverse base station sites.With an even higher degree of coordination,multisite beamforming approaches may be considered. Typically these techniques require a high degree of synchronization and communication between eNodeBs. The use of relay nodes, as shown in the illustration below, is a promising idea to increase the data rates available to edge-of-cell users, or to increase coverage at a given data rate.
Relaying technology has been much studied in academia, resulting in improved understanding of its potential impact on overall system spectral efficiency. Other perhaps more classical approaches may also be considered, such as the support of higher order modulation schemes. This could extend the dynamic range of the link adaptation techniques already available in the first version of LTE. Implementation challenges must also be addressed, though, as ever higher order modulation schemes impose increasingly stringent constraints on the RF components of the transceivers. The challenge for LTE-Advanced will be to deliver these improvements without an unacceptable increase in equipment cost. It is clear that a range of techniques exist which, individually or in various combinations, can be expected to bring significant further improvements beyond the dramatic advances already made by the first version of LTE. These enhancements are likely to be introduced in steps following an evolutionary approach. This is fitting for a system developed by an organization like 3GPP, whose main goal is to define functional systems which are able to evolve gracefully and provide useful service to consumers. The targets for LTE-Advanced are challenging, both in terms of the timescale and expected data rates and performance. To a large extent, theoretical solutions exist. As with LTE, the transition of these solutions from academic theory to real-world networks will be exciting.
