The paradox of 5G is that it has been presented as the biggest sea change yet in mobile communications and yet as a series of evolutionary steps rather than a revolution. That may sound like an oxymoron, but the intention has been to set out a blueprint for mobile communications that will pull innovations through with it and ideally mark the end of clear demarcations between successive generations. It will then only be commercial considerations, a reason for operators and consumers to buy more products, that would drive a stake in the ground for 6G.
Some of these earlier aspirations will be met through the two rounds of 5G standards that are currently being ground out – Release 16, due for release in late 2020, following a few months’ delay as a result of the pandemic; and then Release 17 about a year after that.
The content of both releases can be split into two camps – completely new features and enhancements to functions or capabilities already available in Release 15. The improvements can also be divided by the three primary use case categories (eMBB for higher capacity and bit rate; ultra-reliable low latency communications (URLLC); and eMTC for large numbers of M2M communications that may have only low bit-rates.
Actual use cases may rely on just one of these, two, or even all three. The enhancements can also be analyzed in the light of architectural changes, notably the shift to software-defined operation, and use of network slicing to separate use cases and services, with the ability to apply differential levels of QoS, as well as varying security regimes. There is also increased accuracy over both location and time, as well as improved robustness.
The current Release 15, otherwise known as 5G Phase 1, has set the stage for a number of the new use cases promised by 5G, while failing to deliver all the required capabilities for large-scale deployment across some of the fields. 5G has enabled some of the first Industrial IoT (IIoT) deployments in the sub-6 GHZ spectrum (called FR1), at a broad level across some mines, ports and larger industrial parks. This has galvanized some larger manufacturers in automotive and consumer appliances sectors to apply for their own spectrum licences, especially in Germany.
Volkswagen has applied for 5G spectrum to set up a network inside its factory campuses, while Daimler has applied for a regional licence. Similarly, Siemens and Bosch have applied for spectrum licences in the 3.7-3.8 GHz band to establish their own LTE and 5G networks for ‘digital factories’ under development.
The initial advantage was little more than the security inherent in cellular SIM devices, but low latency combined with higher location accuracy were also motivations. The foundation was laid in Release 15, in the case of low latency with the basic air interface. Then Release 16 has tightened up on other contributors to latency, such as the set-up and activation time associated with CA/DC (carrier aggregation/dual communication), which allows for enhanced bandwidth and load balancing under 5G.
Accurate positioning is also a critical capability of 5G to the extent it could almost be a fourth fundamental use case. This too will evolve through successive 5G releases, with Rel 16 setting the stage by defining how to harness the higher frequencies of the carriers and associated use of massive antenna arrays to obtain the additional degrees of freedom for greater accuracy, compared with LTE. Under Rel 16, the aim is to get positioning accuracy within three meters, which is a significant improvement and adequate for some map-based applications, even without any additional satellite tracking.
But it is not good enough for many of the demanding new target use cases for 5G, including various factory automation applications and also vehicular control in autonomous driving. These will require accuracy within a meter at least, and preferably down to 10 centimeters for some of the robotic assembly and factory picking tasks.
This additional accuracy will come in the first instance by combining the cellular positioning with multi-constellation satellite navigation, using the USA’s GPS, Europe’s Galileo system, Russia’s Glonass and the Chinese BeiDou. Even this will not be quite accurate enough for precision docking applications, which will therefore be enabled by adding localization elements from inertial sensors in vehicles or robots, for example, as well as optical devices. Integration of all such capabilities will be enshrined in successive 5G releases, starting with Rel 17, noting that latency is vital here too. It is no good just knowing accurately where you were a few seconds ago in a self-driving vehicle.
Mission-critical services are another major focus of 5G and the growing role of mobile in the emergency area is discussed separately (see separate item below). We can note here that under Rel 16, mission-critical services generally are being extended to a wider range of applications, beyond public security and civil defence, for which they were developed initially. The aim is threefold: to find new vertical use cases such as railway traffic management or taxi dispatch; to gain reductions in costs through scale economies; and improve robustness through greater exposure in the field.
All of these areas are extended further in Rel 17, which has that clear demarcation between enhancements and novel features, both contributing to the same wider use cases and architectural changes. The Rel 17 package is now cast in stone following conclusion of the 3GPP’s discussions in December 2019. However, as a result of the Covid-19 pandemic, the proposed timeline for Rel 17 roll-out has been put back by about three months, as for Rel 16, with the final stage, including freezing of OpenAPI specifications, now set for December 2021.
Both the enhancements and new features are listed under the three key categories (eMBB, URLCC and mMTC). Under eMBB, a significant enhancement is tuning of multibeam operation in the millimeter wave (FR2) frequencies, in the light of feedback from early commercial deployments. This should improve resiliency for high bit-rate services in dense deployments and is one of the first examples of responding to early experience in the FR2 range above 6 GHz. A more surprising inclusion in the eMBB enhancements column is description of ongoing work on wide area coverage, which does not really qualify as an improvement to a facility already existing.
Under the URLCC column, the Rel 17 enhancement that stands out addresses the IIoT with improved feedback mechanisms for timing synchronization and also for operation in unlicensed bands. This in fact supports the two strongest use cases for ultra-low latency local communication, the other being automotive, especially autonomous transportation. There are also use cases for ultra-low latency that are not localized, such as gaming, but in those cases 5G can only contribute within its own domain and can do nothing about signal transmission delay. In practice, such applications that are highly delay-sensitive are therefore confined regionally.
Under the mMMC category the main Rel 17 enhancement is reduced overhead associated with connection establishment, which is important as this contributes far more to performance for frequent low bit rate messages. This will be valuable for wearables and a multitude of sensors in the field.
When it comes to more radically new features coming on stream with Rel 17, the headlines are support for the new 60 GHz unlicensed band, and tuning of OFDM to conform with the different transmission regulations. Also notable will be support for IP multicast to enable broadcast applications, pitched particularly at emergency services, the IoT and mass software delivery.
Under URLCC, new features are really study items likely only to crystallize in Rel 18, such as evaluating how to support use cases requiring high bit rate and low latency at the same time. This might include applications around XR (Extended Reality), which itself is a study item under Rel 17.
Under mMTC, one focus is support or optimization for mid-tier applications, including wearables, industrial sensors and surveillance cameras, that generate data at rates higher than IoT/LTE-MTC (machine type communication), but lower than full-blown New Radio (NR). The other improvement will be further reduced complexity for set-up and management, as well as power saving or prolongation of battery life. The latter are essential for some embedded devices whose batteries are intended to last for the lifetime of the deployment.
This roadmap gives an idea of 5G as a perpetual journey rather than a single generational change, and technical logic would hold that we stop talking about Gs from now on. Commercial logic, though, plays to a different tune, as does political posturing (see separate item below).