Table 1 Main enablers for the 5G design

D2D communications

System capacity can be increased by a factor of 2 using the same bandwidth when some RRM mechanisms are used for opportunistic access and there exists full cooperation among devices. Latency is also reduced to the order of the TTI length (e.g., 1–2 ms).

With properly selected safety distance, D2D communication can use the spectrum allocated to small cells (mostly uplink band) without affecting the small cell performance.

How to motivate cooperation of users is still one open issue. Battery consumption is something critical in current systems and D2D communication model is only accepted by the end user for him/her own benefit.

Specification is far from being ready for D2D integration. Opposition from some mobile network operators is also an important barrier to overcome.

New air interfaces

Access time can be reduced down to 1.5 ms if new air interfaces are coupled with efficient access procedures (as shown by FBMC evaluations [60]). Due to the very good frequency localization, the transmit power can be concentrated on only very few subcarriers to eventually enhance significantly the expected coverage or to reduce battery consumption. This is of special relevance for machine-type communications, in which payload is very small.

The well-localized signal energy in frequency domain of the multi-carrier signal also allows for efficient access to fragmented spectrum and efficient spectrum sharing, as a minimum amount of guard bands are needed for the signal separation in frequency.

Changing the waveform impacts the signal structures, implying a revolutionary step towards a new radio generation. Backward compatibility thus cannot be guaranteed, however, for a new radio generation this should not be crucial requirement.

Ultra-dense network

Capacity is directly proportional with the number of nodes, provided centralized interference coordination. For indoor cases, the coefficient of proportionality could be as high as 0.73 with ISD of 10 m.

Together with increasing the bandwidth, the use of more nodes is an easy means to achieve desired levels of capacity.

Cost is the main issue of this approach, together with the need for interference coordination. In case of certain isolation between cells, this need for coordination is relaxed.

Traffic concentration

The use of accumulators or concentrators for machine-type communications improves range of coverage and sensors’ throughput. The improvement factor is between two and three thanks to those relays. For the same throughput and coverage needs, traffic concentration reduces battery consumption.

Concentrators also reduce signaling overhead thanks to the coupling of parallel signaling flows.

Performance heavily depends upon the appropriate relay selection.

Moving networks

In vehicles equipped with two access points, one for outside transmission/reception and another for inside users, their best reception chain improves the link budget for the end user up to 9 dB, in cases where the user is outside the car, and up to 24 dB, when the user is within the car. This results in better coverage or higher user throughput, mostly in the cell-edge.

Battery is not a big issue for vehicles, which opens the door for more active collaboration between cars and end-users. Moreover, the number of antennas integrated in vehicles can be much higher than in the handheld devices. This allows for massive-MIMO solutions.

Mobility of access points increase management complexity mostly due to the higher dynamicity of the network.

Localized traffic flows

With a dedicated bandwidth of 80 MHz, end-to-end latency is reduced to 60%, as compared with current LTE-A system. Moreover, half of the traffic can be offloaded from the cellular system.

Use of localized traffic flows is simple to implement and can be easily integrated into current networks.

Universal caching requires storage resources as well as the design of specific signaling mechanisms.

Massive MIMO

Spectral efficiency can be increased by a factor of 20 when using 256 antennas in the transmitter and receiver side as compared with four antenna systems. For the same spatial multiplexing capability as legacy systems (8 streams), beamforming gain reaches 15 dB.

Simple way of increasing cell efficiency mostly for small cell deployments. Fits together with the use of higher frequencies above 6 GHz due to the reduced antenna size.

Pilot contamination is one of the main showstoppers of the use of massive MIMO. Moreover, TDD mode seems to be a must to reduce signaling overhead thanks to the use of channel reciprocity principle. Form factor forces the use of centimeter wave or mmW to compact massive antennas. Finally, performance is very sensitive to mobility and computation burden could make multi-user solutions unaffordable.

Spectrum above 6 GHz

Meeting the 5G user expectations will require much larger bandwidths as today, in the order of 2 GHz, which can only be available at frequencies above 27 GHz (e.g., 27.5 to 29.5 GHz; 40.5 to 42.5 GHz; 47.2 to 50.2 GHz; 57 to 76 GHz; or 81to 86 GHz).

No restrictions on previous allocations of paired FDD spectrum allows for the use of TDD mode, which is much more efficient for channel estimation. Higher frequencies also reduce antenna size, thus permitting massive MIMO implementations.

Path losses are much higher at such high frequencies, which reduces the coverage to small distances or relies on the use of higher order beamforming to overcome such high attenuations.