5G-Advanced: The Next Step of 5G Evolution

Release Date:2021-09-22  Author:By Wang Xinhui  Click:

Status of 5G Standard
Since the first release (Rel-15) of 5G NR was issued in 2019, the global campaign of 5G commercialization has been in full swing. In later releases, apart from further technical enhancements towards traditional eMBB scenarios, the standardization society envisions more vertical applications as the enhancement of 5G. For example, Rel-16 explored more on some emerging use cases like Industrial IoT, V2X, broadcasting and positioning. The on-going Rel-17 is intended to further expand the 5G networks from various aspects: to further optimize the IoT features for vertical industry applications; to support more types of devices and applications, such as wearables, video surveillance, industrial wireless sensors, and XR devices; to explore more spectrum at even higher mmWave frequency bands from 52.6 to 71 GHz; and to support new network topologies such as non-terrestrial network, integrated access and backhaul, and sidelink transmission.

Emerging of 5G-Advanced
With the massive deployment of 5G base stations and widespread use of 5G cell phones, people are beginning to think about the future, i.e. 6G. Recently, IMT-2030 disclosed a white paper for 6G vision and potential techniques, in which some popular directions and use scenarios have been identified, such as immersive cloud XR, integrated terrestrial and non-terrestrial networks, integrated sensing and communications and holographic communication.
From the standardization point of view, the first release of 6G specification from 3GPP is likely to be ready by 2029-2030. Before that, there will be another three or four releases for the evolution of 5G. In the recent 3GPP/PCG#46-e meeting held in April 2021, it has been agreed that "5G-Advanced" will be used to identify 3GPP specifications and reports from Release 18 onwards. With the new marker adopted mainly for the marketing interests towards 2025, it is important to keep digging the potentials of 5G network to continue the prosperity of 5G industry. Actually, many of the "6G techniques" under hot discussion can be built on 5G framework, and thus it would be better to also include them in the visions of 5G-Advanced as long as the use cases are clear and the techniques are expected to be mature enough in the next few years.

Key Techniques for 5G-Advanced
The 3GPP wireless standards evolve continuously and gradually. In every new release, there will be some continuous work either for the left-over issues or for further enhancements on the past release. It is foreseen that the natural overlapping exists as well between 5G stage-I and 5G-Advanced, or between 5G-Advanced and 6G. On the one hand, some continuous work from Rel-17 can be expected to be further specified in Rel-18. On the other hand, the 5G NR should be further extended to support more scenarios and use cases that have real market need by 2025, in particular for IoTs and verticals from different dimensions. Those potential enhancements are shown in Fig. 1, where some new requirements in terms of the combination of KPIs are considered.

Intrinsic AI
AI is expected to be the core functionality with standard support in 5G advanced networks. The meaning of "intrinsic AI" is reflected on two folds, one is "AI for 5G advanced" which means the design of AI to enhance communication systems, and the other is "5G advanced for AI" which means the design of systems should also be applicable for better support of AI.
In Rel-17, 3GPP picked three most popular use cases to be studied firstly: network energy saving, which enables smart switch on/off of certain base stations or certain carriers according to the traffic tide prediction and scenario recognition; load balance, which enables ML model based load prediction to improve load balancing performance; and mobility optimization, as many radio resource management actions related to mobility can benefit from the predicted UE location/trajectory.
Other than the network aspects, AI applications to the physical layer processing are much more challenging, especially for the fundamental channel coding and modulation. Nevertheless, for some procedure-wise designs with the need of prediction or estimation, AI can be applied to avoid potential mismatch of information and therefore enhance the system performance, for example for the beam management or CSI compression. In short, although the physical layer applications are quite challenging, it is definitely the right time to study those possibilities in the 5G-Advanced time window, starting from identifying the typical use cases, the system models, and the corresponding evaluation methodology in Rel-18.
Ubiquitous Network
FR2 at millimeter wave band is a key enabler for 5G-Advanced, which can unlock wider bandwidth, higher throughput, and lower latency compared with FR1 operating in sub-6 GHz. Although FR2 is supported from the beginning of 5G NR, there are still some challenges for the FR2 deployment, e.g. coverage limits due to blockage, frequent handover and inter-cell interference due to the ultra-dense small cells. Further improvements in terms of network robustness are needed to make it more widespread worldwide in the 5G-Advanced timeline.
Reconfigurable intelligent surfaces (RIS) is a promising technique for the future wireless network, especially when the frequency range goes higher and the coverage is the key issue to be solved. As an energy-efficient and cost-effective solution, RIS can be applied to effectively control the wavefront (e.g. phase, amplitude, frequency, and even polarization) of the impinging signals in an active or passive way. Appropriate deployment of RIS can be used to provide reliable non-line-of-sight propagation when the line-of-sight path is blocked, or to create either beam-formed signals to the target users or beam-nulled signals to the interferers, and therefore to improve the throughput and energy efficiency for the cell-edge UEs. It is envisioned that RIS would be a good candidate for 5G-Advanced deployment based on the existing 5G framework. In long run, RIS technique could go further to build up smart network paradigm, based on a joint optimization taking into account not only the transceiver design but also the wireless environment. In addition, multiple connectivity can be used to improve the robustness which can be fulfilled basically by the extension of existing dual-connectivity.
NR Based Massive IoTs
Although it has been claimed that the requirements of mMTC can be fulfilled by LTE NB-IoT/eMTC with the original individual KPIs of connection density, coverage and power consumption, it may not be suitable for the modern MTC scenario, for example, the use case of wide area sensor monitoring and event driven alarms which requires not only massive connections but also highly efficient data transmission; and also the smart grid use cases, where the advanced smart metering requires very high connection density with good coverage, decent data rate, low latency and high reliability. Similar requirement on both massive connection and certain URLLC requirements can be found in the modern factories, such as condition monitoring for safety, packaging machine, process automation, motion control, mobile robots, etc.
Compared with the LTE based solutions, massive IoT based on NR is more attractive because of the superiority on the availability of more spectrum, flexibility of scheduling/allocation, and additional spatial domain resources given by beam based operation. Grant-free non-orthogonal multiple access (NOMA) is a promising technique for NR-based massive IoT in 5G-Advanced. On the one hand, grant-free allows uplink transmission with lower latency and more energy efficiency, because the random access procedure and data transmission are completed simultaneously. On the other hand, NOMA is a perfect solution to cope with the potential collisions among multiple UEs performing grant-free transmission, so as to achieve highly efficient transmission for massive devices without loss of reliability. In addition, higher layer optimization such as intelligent distribution of UEs in different RRC states can be studied to minimize the need for load triggered handover and therefore improve the transmission efficiency and connection density.
Industrial IoT 2.0
For IIoT 2.0, it is mainly concerned about the application of UL heavy scenarios. With some extreme use cases such as the machine vision applications in the modern factory, or the broadband access in a crowd at stadiums or concerts where the users want to share what they see or they hear, a higher throughput requirement is put on the uplink than the downlink. In addition to a very high connection density, the required uplink data rate could be in the order of Gbps or 10Gbps with quite low latency requirement, which can be hardly fulfilled by the current NR design. 
One possible way is to support enhanced features based on the existing NR design, such as higher number of antennas or MIMO layers. However, these enhancements require the initiating device for the uplink transmission to be powerful enough, while usually the capability of a UE is quite limited compared to a base station. The idea of boosting the UE capability can be realized through user virtualization and cooperation for the devices in proximity or those belonging to the same owner. Multiple devices can form a virtual user by sharing their capability of MIMO, carrier, etc., and collaboratively transmit and receive data from the network to enhance the quality of transmission.
Another bottleneck for the UL heavy applications is about the traffic congestion. There could be more fundamental enhancements based on the idea of MESH network. For example in modern factory, many of the communications are actually local between the proximate devices. In such case, Mesh based networking can be helpful to offload the traffic between proximate devices to sidelink, so as to alleviate the traffic load going through the core network. Those design aspects include intelligent routing path discovery, latency/load aware routing, etc. 
Blockchain has the ability to achieve secure, immutable and decentralized data storage with low latency in 5G-Advanced, enabling trustworthy network sharing among the operators. 
Some key data, such as cell resource or spectrum utilization, can be uploaded to the Blockchain platform through the sharing base stations of the hosting operator. The Blockchain platform is used to guarantee the trustworthiness and transparency of network sharing, and avoid disputes among operators. Other participating operators or non-operator participants can access to the Blockchain platform to get the trusted data according to the operators' policy and regulation needs. In addition, the hosting operator can optimize the allocation of system resources in a high-efficiency mode based on the operators' resource occupancy from the Blockchain platform. The host operator is responsible for providing services to the customers of participating operators in network sharing, and reporting the faults from those customers to the Blockchain platform to prove they are treated equally as the host operator's customers. Participating operators can acquire the fault reporting of their customers through the Blockchain platform. With that, a secure and trustable network operation can be achieved.

Conclusion Remarks
There is no doubt that 5G-Advanced will be the next era to continue the prosperity of communication industry and to bridge the gap from 6G vision. ZTE is willing to work together with our industry partners, to jointly build a safe, intelligent, and sustainable 5G-Advanced standardized industry.