Envisioning and Defining 6G Together

”Ten years may sound like a long time, but it passes very quickly. Whether or not we will provide a

satisfactory answer by 2030 will depend on some crucial factors: Was the process of defining 6G

open? Was there broad participation from a diverse range of players? Was there sufficient

engagement? Have we delivered an attractive 6G vision? ... We hope that it will inspire more

people, companies, and industries to bring broader and deeper perspectives to 6G. Huawei is also

ready and willing to engage with our industry peers, with industry verticals, and with enterprises

that may need 6G. Let's envision and define 6G together.”

Eric Xu

Deputy Chairman of the Board and Rotating Chairman of Huawei

 Content

2 1. Mega-trends and Key Drivers

17 4. New Elements
Driver 1: New Applications & New Business New Spectrum
Driver 2: Proliferation of Intelligence New Materials and Antennas
Driver 3: Sustainability & Social Responsibility New Channels
New Devices

3 2. Overall Vision and Capabilities

23 5. Enabling Technologies
and Architectures
Pillar 1: Native AI
Pillar 2: Networked Sensing Paradigm Shifts in the Air Interface Design
Pillar 3: Extreme Connectivity Paradigm Shifts in the Network Architecture Design
Pillar 4: Integrated NTN
Pillar 5: Native Trustworthiness
Pillar 6: Sustainability

10 3. Use Cases and Requirements

27 6. Summary and Roadmap
Usage Scenario 1: eMBB+ Summary of Paradigm Shifts
Usage Scenario 2: URLLC+ 6G Roadmap
Usage Scenario 3: mMTC+
Usage Scenario 4: Sensing
Usage Scenario 5: AI
 Table of Figures

Figure 1 Mega-trends of mobile communications toward 2030 and beyond 2

Figure 2 6G features driven by proliferation of intelligence 3

Figure 3 Fusion of physical, biological, and cyber worlds 4

Figure 4 Six pillars of 6G key capabilities 5

Figure 5 AI4NET and NET4AI 5

Figure 6 Networked sensing enables new services beyond communication 6

Figure 7 RAN KPIs for extreme connectivity 7

Figure 8 Integrated NTN for low-latency long-distance communication 8

Figure 9 Dimensions of trustworthiness and the multilateral trust model 9

Figure 10 Two typical challenges for low power consumption in the 6G era 10

Figure 11 Overview of typical 6G use cases 10

Figure 12 Multilayered frequency band framework 18

Figure 13 Use case example of RISs enabled by reconfigurable materials 19

Figure 14 New challenges in channel modeling for 6G 20

Figure 15 Capabilities and trends for future devices 21

Figure 16 Paradigm shifts of future air interface design 23

Figure 17 Paradigm shifts of future network architecture design 25

Figure 18 Summary of paradigm shifts in 6G 27

Figure 19 Expected timeline for 6G standardization 28

 6G: The Next Horizon
From Connected People and Things to Connected Intelligence
Huawei Technologies Co., Ltd.

Abstract

Wireless communication turned its first page in the early 1900s when Marconi transmitted the radio signal across the
Atlantic. Since the 1980s, mobile communication has revolutionized the world, transforming every aspect of our lives.
With the endless frontiers spanning 5G, we start wondering what 6G will be like. 6G — a more advanced next-generation
mobile communication system — will go far beyond just communications. It will serve as a distributed neural network that
provides links with integrated communication, sensing, and computing capabilities to fuse the physical, biological, and
cyber worlds, ushering in an era of true Intelligence of Everything. Building upon 5G, 6G will continue the transformation
from connected people and things to connected intelligence. In essence, it will bring intelligence to every person, home,
and business, leading to a new horizon of innovations. In this paper, we present a holistic view of our 6G vision, exploring
6G key capabilities, new use cases and requirements, new building blocks, and paradigm shifts in air interface and network
architecture designs. More details are available in our book "6G: The Next Horizon" recently published by Cambridge
University Press [1].

Keywords

6G, connected intelligence, native AI, networked sensing, integrated sensing and communication (ISAC), extreme
connectivity, integrated terrestrial and non-terrestrial networks, native trustworthiness, sustainability, paradigm shifts

1 | Huawei Technologies
Technology
1. Mega-trends and Key Drivers At the same time, with the burgeoning numbers of
IoT devices and the new capability of wireless sensing
providing big data to learning algorithms, AI will become an
New generations of mobile communications sys tem
engine for all types of automation. Big data will therefore
emerge roughly every 10 years, while the mainstream
become a major driver for the order-of-magnitude
services provided by mobile networks and the application
increase in 6G network throughput. Furthermore, high-
of new frequency bands usually take two generations or
per formance indus trial IoT applications will impose
more to mature. In fact, it took almost four generations
demanding requirements in terms of deterministic latency
to have pe ople conne c te d any whe re and any t ime,
and jitter, while also needing guaranteed availability and
leading to a connected society. With the rapid global
reliability. Such use cases also drive the extreme and
commercialization of 5G s tar ting around 2020, not
diverse performance that will be a defining feature in 6G.
only will society be better connected with enhanced
communication capabilities, but also more devices in all
kinds of business scenarios will be connected, moving Driver 2: Proliferation of Intelligence
from an era of connected society to one of connected
everything. Following this trend, we envision that 6G will The mobile industry has profoundly impacted people's
provide better connections for people and things, and life, helped to mitigate the digital divide, and contributed
will embrace the trend of a smart society, continuing significantly to society's overall productivity and economic
the transformation from connected people and things growth. This trend will continue into 2030 and beyond.
to connected intelligence. In addition to the ongoing In particular, as pervasive intelligence — supported by
evolution of the three usage scenarios initiated in 5G [2], massive machine learning (ML), brute-force computing,
AI and sensing will become two new usage scenarios in and big data analytics — becomes the key enabler of
6G, as suggested in Figure 1. Three key drivers are leading business and economic models in the future [3], paradigm
mobile communications toward a new era of connected shifts in radio technology and network architecture will be
intelligence, as described below. driven by the following four critical factors, as shown in
Figure 2.

Driver 1: New Applications & New

Business • ^¼ŔąűãD…řŅŅīƑ

For 6G, end-to-end (E2E) mobile communications systems

In the 6G era, more applications will emerge. Extended
will be designed with optimal suppor t for AI and ML
reality (XR) cloud services together with haptic feedback
— not only as a basic functionality, by also for optimal
and holographic display are likely to become mainstream
efficiency. In terms of the architecture, running distributed
human-centric applications. The exponential increase in
AI at the edge will achieve ultimate performance while
the traffic demand per device, together with strict latency
also addressing the concerns of data ownership. Truly
and reliability requirements, will become a major challenge
pervasive intelligence, combined with deeply converged
for 6G network design in terms of the massive capacity
ICT systems that feature diverse connectivity, computing,
needed.

Figure 1 Mega-trends of mobile communications toward 2030 and beyond

Huawei
HuaweiTechnologies
Technology | 2
 Figure 2 6G features driven by proliferation of intelligence

and storage resources at the edge, will become a native collaboration and convergence of ICT and OT sectors
trait. The 6G network architecture with native AI support — would be extremely beneficial. The first wave of 6G
will bring "Networked AI", moving away from today's commercial use is likely to boost both the consumer and
centralized "Cloud AI" [3]. vertical markets.

• Native Data Protection Driver 3: Sustainability & Social

Responsibility
The protection of privacy in every aspect of 6G networking
and data will be essential. We expect that users — which
With multiple generations of technologies and spectrum
might be people or machines — will be empowered as data
deployment s coexis ting in mobile sys tems and the
owners with control and operation rights. At the same
increasingly heterogeneous ser vices running on top,
time, the design of 6G should guarantee privacy, ensure
there is high demand for sustainable development of
the proper rights of data subjects, enable data control
6G innovations. It is expected that deploying, operating,
and processing, and support policies such as the General
monitoring, and managing 6G networks and ser vices
Data Protection Regulation (GDPR) [4] in order to establish
will be cost- and energy-efficient, easy, and automated.
fundamental guidelines for technology design and usage
Furthermore, 6G should be an enabler for achieving the
in the future.
sustainable development goals (SDGs) [7] of society as a
whole.
• ^¼ŔąűãŒňřŌŔŲīƑāąģãŌŌ

To support a diverse range of use cases and markets, it is

2. Overall Vision and Capabilities
essential to have customized, verifiable, and measurable
trustworthiness. Today's nomothetic network ownership 6 G — a m o r e a d v a n c e d n e x t- g e n e r a t i o n m o b i l e
and operation will evolve into a many-party, many-player, communic at ions s ys te m — w ill go far beyond jus t
and many-actor pattern, where an inclusive multilateral communications. Over the next decade, in addition to
trust model — rather than a unilateral one — will be vital. continuous wireless innovations, the rise of massive AI and
In addition to being future-oriented, the trustworthiness the creation of massive digital twins will be the two major
a rc h i te c ture s h o ul d in c o r p o rate s e c ur i t y, p r i va c y, catalysts that fuel more technology breakthroughs. The
resilience, safety, and reliability [5]. resulting 6G will be a game-changer in terms of both the
economy and society — it will lay a solid foundation for the
future Intelligence of Everything.
• ^¼Ŕąűã%ąűãňŌąƍãß)ÙīŌŸŌŔãġ
We envision that 6G will enable the transformation from
With 5G capabilities gradually expanding, the vertical connected people and things to connected intelligence.
wireless market is expec ted to ramp up throughout Compared with its predecessor, 6G will offer extreme
the 2020s. As we approach the 6G era, a universal performance and realize major improvements in terms
ICT framework — one that could offer an overarching of key per formance indicators (KPIs). Fur thermore,
perspective for all industries and thereby accelerate the it will be a key e nable r in achieving t he f ull-s c ale

3 | Huawei Technologies
Technology
 Figure 3 Fusion of physical, biological, and cyber worlds

digital transformation of all vertical businesses. More Pillar 1: Native AI

importantly, we believe that 6G will serve as a distributed
neural network, providing links with integrated sensing, 6G will boast a native AI capability, which is neither
communication, and computing capabilities. This will fuse an add-on nor an over-the-top feature. One of the
the physical, biological, and cyber worlds, ushering in an primary objectives for 6G is to support AI everywhere.
era where everything will truly be sensed, connected, and AI will be both a service and a native feature in the 6G
intelligent. communication system, and 6G will be an E2E system that
supports AI-based services and applications. Specifically,
Figure 3 provides an overview of how the 6G network
6G air interface and network designs will leverage E2E
functions as the fabric of the converged physical and
AI and ML to implement customized optimization and
cyber worlds. First, from the cyber world to the physical
automated operation, administration, and management
world — this is a typical downlink channel — the primary
(OA&M). This is known as "AI for Network (AI4NET)", as
service will be all kinds of XR, which, enhanced by tactile
shown in the upper half of Figure 5. In addition, each 6G
communication and new human-machine inter face,
network element will natively integrate communication,
creates immersive experience when interacting with the
computing, and sensing capabilities, facilitating the
digital world. Meanwhile, the continuous deep learning
evolution from centralized intelligence in the cloud to
in the digital world serves as AI engine for the physical
ubiquitous intelligence on deep edges. This is the concept
world, providing real-time inferences to facilitate all
of "Network for AI (NET4AI)" or "AI as a Service (AIaaS)", as
kinds of decision making. This imposes major challenges
shown in the lower half of Figure 5. For AIaaS, 6G functions
on the radio interface design, requiring extremely high
as a native intelligent architecture that deeply integrates
throughput and ultra-low latency. Second, from the
communication, information, and data technologies,
physical world to the cyber world — this is a typical
as well as industry intelligence, into wireless networks,
uplink channel — the primary application is sensing and
ser ving all types of AI applications with large-scale
the collection of big data for ML. How to compress and
distributed training, real-time edge inference, and native
transmit the huge amount of data (or model parameters
data desensitization.
in the neural networks) for massive ML poses a new
challenge. 6G will be a network of sensors and ML, where To achieve this vision, it is necessary to address three key
data centers will become neural centers, and ML tasks will challenges.
spread over the entire network, from neural center to deep
neural edges (e.g., base stations or even mobile devices). • 6G should be the most efficient platform for AI.
This presents new challenges in terms of how to
In the following sections, we discuss the six fundamental realize minimum cost for both communication and
technology pillars that will shape 6G, that is, the six computation, each of which is a KPI for future study.
technical directions for 6G, as shown in Figure 4. To minimize communication costs, it is necessary to
design a 6G system that can transfer massive big

Huawei
HuaweiTechnologies
Technology | 4
 Figure 4 Six pillars of 6G key capabilities

Figure 5 AI4NET and NET4AI

data for AI training using minimal capacity resources. • An efficient and distributed collaborative learning
To minimize computation costs, it is necessary to architecture will be vital for reducing the computational
implement optimally distributed computing in the load involved in large-scale AI training. Data split
networks, where we can best leverage mobile edge and model split for AI will be incorporated into the
computing. 6G network architecture. Furthermore, leveraging
distributed and federated learning will help optimize
• In order to support ML, 6G will need to enable the computing resources, local learning, and global
collection of massive data from the physical world learning, and help meet the new data local governance
(millions of times more data than at present) so requirements. In this sense, 6G core network functions
that a cyber world can be created. This, however, will be pushed toward a deep-edge network, while
poses another major challenge for 6G. As such, cloud-based software operations will shift toward
how to effectively compress training data based on massive ML. In addition, with the frequent transfer of
information and learning theory becomes a new and large amounts of data and models from deep edges
essential topic in 6G research. (devices), the 6G radio access network (RAN) will shift
from downlink-centric to uplink-centric.

5 | Huawei Technologies
Technology
 Figure 6^ãŔŲīňėãßŌãģŌąģùãģ¼ØěãŌģãŲŌãƒąÙãŌØãŸīģßÙīġġřģąÙ¼Ŕąīģ

Pillar 2: Networked Sensing and communication (ISAC) design in the 6G network has
two targets and potential benefits: to significantly reduce
6G will feature the networked sensing capability. Higher the cost of additional sensing equipment, and to leverage
fre quenc y bands (from mmWave up to THz), wider the large-scale cooperation between widely deployed
bandwidth, and denser distribution of massive antenna base stations and user devices for improved sensing
arrays in future 6G systems will enable a single system to performance.
integrate wireless signal sensing and communication, each
The ISAC functions can happen at different levels, ranging
of which mutually enhancing the other. The communication
from loosely coupled to fully integrated and from shared
system as a whole can serve as a sensor, exploring radio
spectrum and hardware to shared signal processing and
wave transmission, reflection, and scattering in order
protocol stacks. It can even include cross-module, cross-
to sense and bet ter understand the physical world,
layer information sharing. Such integration will bring
ultimately providing a broad range of new services. This
mutual benefits. Furthermore, it will enable technological
is known as "Network as a Sensor". Four categories of use
innovations on new system KPIs and fundamental limits,
cases that can be supported by 6G sensing are shown
new channel model and evaluation methodologies, joint
in Figure 6 and described later in Section 3. In terms of
waveform design, hardware co-design, new frameworks
sensing, it enables high-accuracy localization, imaging,
of protocols and procedures, cooperative sensing and
and environment reconstruction capabilities that could
data fusion, AI-assisted sensing, sensing-assisted ML, and
help improve communication performance — for example,
much more.
more accurate beamforming, faster beam failure recovery,
and less overhead to track the channel state information
It is also worth mentioning that recent developments in
(CSI). This is known as "sensing-assisted communication".
semiconductor technology have bridged the "Terahertz
Moreover, as a foundational feature for 6G, sensing is
(THz) band gap" (caused by the lack of THz hardware
a "new channel" that observes, samples, and links the
enablers). These developments are expected to stimulate
physical and biological worlds to the cyber world. Real-
various THz sensing applications [8]. In addition to ultra-
time sensing is therefore essential to make the concept of
high resolution imaging, given the range of wavelengths
digital twin — a true and real-time replica of the physical
and properties of molecular vibration, THz sensing can
world — a reality in the future.
perform spectrogram analysis to identify the constituent
par t s of dif ferent t ypes of food, medicine, and air
Traditionally, sensing is a standalone function with a set
pollution. Furthermore, due to its compact form-factor
of dedicated devices and equipment, such as radar, lidar,
and non-ionizing safety, THz sensing can be integrated
computed tomography (CT), and magnetic resonance
into mobile devices and even wearables to identify the
imaging (MRI). Mobile phone positioning in mobile
number of calories in food and help detect hidden objects.
systems, assisted by air interface signaling and device-
As a result, 6G sensing devices will become a gateway for
based measurements, is an elementar y sensing-like
realizing numerous innovative AI applications.
capability. Compared with the traditional methods of
providing sensing functionality, the integrated sensing

Huawei
HuaweiTechnologies
Technology | 6
 Figure 7 RAN KPIs for extreme connectivity

~ąěě¼ňɥ˜)ŷŔňãġãīģģãÙŔąűąŔŸ Communic ation through visible light is a potential

radiation-free transmission technology that enables
6G will provide universal high-per formance wireless connectivity without an electromagnetic field (EMF).
c on n e c t ion s a n d ul t imate ex p e r ie n c e w i t h s p e e d s However, a large-scale micro-LED array technology will
comparable to optical fiber. The major KPIs for the 6G RAN be required to attain data rates reaching tens of Tbit/s
are shown in Figure 7. Up to Tbit/s peak rate, 10–100 Gbit/s for shor t-dis tance communications with low power
experienced rate, sub-millisecond level latency, a tenfold consumption, small form factors, and low-cost devices.
increase in the density of 5G connections, centimeter- In addition, visible light communication (VLC) [10] can
level localization, millimeter-level imaging, and E2E system access large amounts of unlicensed spectrum, but several
reliability based on controllable error distribution will challenges in terms of uplink transmission, mobilit y
not only enable human-centric immersive services in the management, and high-performance transceivers must be
future, but also accelerate full-scale digital transformation overcome before VLC can be successfully utilized in 6G.
and productivity upgrade of vertical industries.

Pillar 4: Integrated NTN

The increase in wireless traffic drives the need for wider
spectrum that usually comes with higher frequency, while
6G will integrate terrestrial networks and non-terrestrial
the mobile communications system infrastructure favors
networks (NTNs) to deliver complete coverage worldwide,
the lower frequency spectrum for ubiquitous coverage.
connecting the unconnected. As the cost to manufacture
After several generations of wireless evolution, more and
and launch satellites decreases, huge fleets of low- or very
more new frequency bands are deployed for network
low-earth orbit (LEO/VLEO) satellites will become reality
upgrades. 6G will use not only the millimeter wave
in NTNs — it is a strong possibility that 6G will integrate
(mmWave) spectrum, but also the THz or even visible light
VLEO satellite mega constellations. A VLEO satellite
spectrum, potentially using the entire spectrum for the
system, in addition to delivering worldwide coverage,
first time in order to deliver ultimate extreme connectivity.
offers a number of new capabilities and advantages.
For example, it eliminates the issue with communication
THz communication is a new wireless technology that
latency inherent in conventional geostationary earth orbit
involves numerous challenges. Research is currently
(GEO) and medium earth orbit (MEO) satellite systems.
exploring the design of high-power devices, new materials
It can also provide coverage to areas uncovered by
for antennas, radio frequency power transistors, THz
terrestrial networks, offering complementary radio access.
transceiver on-die architecture, channel modeling, and
VLEO satellite systems can also provide more accurate
array signal processing [9]. Whether THz technology is
positioning, which is critical for autonomous driving and
successfully adopted in 6G depends on the engineering
important for earth sensing and imaging. Figure 8 shows
break throughs in THz-related component s such as
some preliminary region characterization where using
electronic, photonic, and hybrid transceivers and on-die
mega-LEO satellite constellation would achieve lower
antenna arrays.

7 | Huawei Technologies
Technology
 Figure 8 Integrated NTN for low-latency long-distance communication

transmission latency than traditional fiber over long- Data, as well as the knowledge and intelligence derived
distance communication. from it, is the driving force behind 6G network architecture
redesign, wherein new features will be developed to
In addition to satellite communications, new radio nodes enable E2E native trustworthiness. These features include
such as drones, unmanned aerial vehicles (UAVs), and high- new data governance architectures that support data
altitude platform stations (HAPSs) will be an integral part compliance and monetization, as well as advanced privacy
of 6G, functioning as either mobile terminals or temporary protection and quantum attack defense technologies.
infrastructure nodes. By integrating both terrestrial and
non-terrestrial networks, 6G will stand apart from its From a technology perspective, the security, privacy,
predecessors. and resilience established by cryptography and defense
technologies are usually referred to as the three pillars
NTNs are currently designed and operated separately. of trustworthiness, which are underpinned by ten blocks
In the 6G era, however, their functions and operations, (three in security, two in privacy, and five in resilience), as
along with their resources and mobility management, shown in Figure 9. The design objectives with regard to the
are expected to be tightly integrated. Such an integrated three pillars and ten blocks are summarized as follows:
system will identify each user terminal with a unique ID,
unify billing processes, and continuously provide high- • Balanced security: Different protected assets or
quality services via optimal access points. Moreover, with properties may require a different level of protection
a virtualized air interface, the addition and deletion of a o r d i f fe re n t we i g h t i n e a c h f a c e t of i n te g r i t y,
non-terrestrial access point would be transparent to user conf ide nt iali t y, and availabili t y, de p e nding on
equipment (UE). Given that the deployment, maintenance, different scenarios.
and energy source of satellites differ completely from
those of terrestrial networks, it is expected that new • ~ãňġ¼ģãģŔ Ņňąű¼ÙŸ ŅňīŔãÙŔąīģ˜ The identity and
operating and business models will emerge. behavior of users are protected so that only those
parties authorized by users are able to interpret the
content of information transferred among them.
~ąěě¼ňɧ˜^¼ŔąűãŒňřŌŔŲīƑāąģãŌŌ
• …ġ¼ňŔ ňãŌąěąãģÙ㘠In order to provide and maintain
The 6G network will integrate various capabilities such an acceptable level of service while operations face
as communication, sensing, computing, and intelligence, various faults and challenges, situation awareness
making it necessary to redefine the network architecture. and big data analytics are leveraged to identify and
The novel network architecture should be capable of being then avoid or transfer risks. If this is not possible,
flexibly adapted for tasks such as collaborative sensing the consequences must be controlled and only the
and distributed learning to proliferate AI applications on a residual non-harmful risks accepted [11].
large scale, where trustworthiness should be guaranteed
as a native feature. The concept of "trustworthiness" here Among the enabling technologies, the following two are of
covers topics including security, privacy, resilience, safety, note:
and reliability [5].

Huawei
HuaweiTechnologies
Technology | 8
 Figure 9%ąġãģŌąīģŌīøŔňřŌŔŲīƑāąģãŌŌ¼ģßŔāãġřěŔąě¼Ŕãň¼ěŔňřŌŔġīßãě

• \řěŔąě¼Ŕãň¼ě ŔňřŌŔ ġīßãě˜ An inclusive multilateral energy efficiency (defined in bits per Joule) 100-fold
t r u s t m o d e l (i n c l u d i n g m o d e s s u c h a s b r i d g e, across the network and keep the total energy consumption
consensus, and endorsement) will ser ve as the (in unit of Joules) lower than 5G while also ensuring
foundation of future security systems. Because the 6G optimal service performance and experience. As the core
network architecture will trend toward a distributed infrastructure of the digital economy, 6G will have to make
nature, a consensus-based model may be the most unique contributions to the sustainable development of
important mode in the multilateral trust model. For humankind.
this purpose, distributed ledger technologies (such as
blockchain-like technologies) will be developed after In terms of the research directions for E2E green 6G
new challenges in wireless networks are addressed. network design, the potential technologies to realize
Such challenges center on how to achieve low latency, energy efficiency span architectures, materials, hardware
high availabilit y, high reliabilit y, s trong privac y compon e nt s , algor i t hm s , s of t ware, an d protocol s .
protection, and digital sovereignty. Indus tr y consensus needs to be es tablished on the
methodology used to evaluate sustainability across the
• ~ī Ō Ŕ̀ Ň ř¼ ģ Ŕř ġ Ù ň Ÿ Ņ Ŕī ù ň¼ Ņ āŸ ˜ A s q ua ntum entire ecosystem. Dense network deployment (leading
computing continues to develop, challenges arise to a shor ter propagation distance), centralized R AN
with regard to classical cryptography, which is based architec ture (resulting in fewer cell sites and higher
on mathematic al problems such as large prime resource efficiency), energy-aware protocol design, and
factorization and discrete algorithms. Key generation cooperation between users and base stations are some
and exchange algorithms are two indispensable factors that need to be carefully considered in order to
elements involved in cryptography. In 6G, one-time achieve an energy-efficient 6G communication system.
pad (OTP) encryption can be used with full-duplex In addition, renewable energy and radio frequency (RF)
communications at the physical layer in order to energy harvesting technologies as well as backscattering
safeguard against quantum computing–based attacks. communication techniques (with no active RF power)
When quantum computing becomes reality, quantum should also be considered. As we move toward using
communication technologies are expec ted to be higher and higher frequencies, finding innovative ways
more secure and have lower latency due to quantum to deal with the reduced power amplifier (PA) efficiency
entanglement. Lightweight cryptographic algorithms becomes a major challenge [12], as shown in Figure 10-a.
and privacy-compliance-related algorithms are some
potential areas in this regard that warrant further Another significant challenge centers on computing power
research. consumption due to the rise of AI. We can speculate
that, on average, the human brain achieves data rates of
20,000 Tbit/s and can store 200 TB of information while
Pillar 6: Sustainability consuming only 20 Watts. Conversely, the computing
power of AI is doubling every two or three months, far
Green and sus tainable development is the core in excess of Moore's Law. For a neural center to achieve
requirement and ultimate goal of network and terminal the same capabilities as the human brain, there is a 1,000
designs in 6G. By introducing the green design concept times gap at a point of time near the end of Moore's
and native AI capability, 6G aims to improve the overall

9 | Huawei Technologies
Technology
 Figure 10 Two typical challenges for low power consumption in the 6G era

Law, as shown in Figure 10-b. In order for neural centers Among these categories, eMBB+, URLLC+, and mMTC+
to replace data centers and fully leverage the potential are extensions and combinations of the usage scenarios
of AI, it is imperative to use significantly advanced ML defined in 5G, while sensing and AI are two new usage
technologies that facilitate sustainable AI-based 6G [13]. scenarios that will flourish in 6G. In the following section,
A standardized approach to implementing a distributed we explore these categories and provide examples of use
computing architecture and software orchestration will cases and requirements under each category.
enable the 6G network to be an efficient platform for a
diversified ecosystem.
Usage Scenario 1: eMBB+

ɥ˚‘Ōã¼ŌãŌ¼ģ߁ãŇřąňãġãģŔŌ This usage scenario is the continuous evolution of

enhanced mobile broadband (eMBB) for human-centric

There is a tendency to overestimate what can be done communic ation use c ases . It will enable ex tremely

in two years but underestimate what can be done in ten immersive experience and multi-sensory interactions in XR

years. As new technologies become more widely adopted applications — including augmented reality (AR), virtual

in wireless communications systems, within the lifecycle of reality (VR), and mixed reality (MR) — and telepresence.

6G, many aspects of our daily lives will be augmented by eMBB+ will pose much higher requirements on the peak

ultra-high speed and ultra-reliable wireless connections, data rate, user-experienced data rate, low E2E latency, and

native AI, and advanced sensing technologies. Based on large system capacity (i.e., high throughput and supported

the key capabilities required, we have identified five major connections). Furthermore, it will enable a range of use

categories of usage scenarios, as shown in Figure 11. cases in entertainment, education, manufacturing, and

Figure 11dűãƒąãŲīøŔŸŅąÙ¼ěɨ<řŌãÙ¼ŌãŌ

Huawei
HuaweiTechnologies
Technology | 10
 navigation, transforming the way we live, learn, work, and adjust their operation time, stress, gesture, and so on. This
travel. Both indoor and outdoor cases are needed, where type of interactive teleoperation requires very low latency,
seamless user experience in the target activity areas with the RTT requirement for air interface transmission
must be guaranteed along the E2E routes of activities, being as low as 0.1 ms. Fur thermore, teleoperation
regardless of the high mobility in extreme cases. The user- imposes strict requirements on the relative transmission
experienced data rate in remote areas and on planes and latency between audio, video, and haptic information, as
ships must be maintained to support ubiquitous high- well as on reliability and throughput.
quality connections. Some examples of the corresponding
use cases are discussed below.
• Glass-free 3D and holographic
displays
• ‘ěŔąġ¼ŔãąġġãňŌąűãÙěīřß©
While wearing VR devices, users always focus on the
360 o extremely immersive XR is an evolution of current
screen regardless of whether the displayed object is close
XR services, offering an even higher resolution and video
or far away. Because this affects users' ability to perceive
frame rate close to the limit of human perception. It
depth correctly, they may experience dizziness or other
provides an extremely low interactive latency, delivering
unwanted effects. Glass-free 3D displays based on visual
the optimal immersive visual experience. For example, it
accommodation are expec ted to be the nex t game-
will enable people to play football virtually with friends
changing solution, relying on techniques such as light field
anywhere and at any time, or watch a live football match
and holographic display. Such displays would allow users
from the referee's perspective. To enable extended periods
to see far-away family members up close without the need
of use without making users experience dizziness, motion
to wear glasses, delivering an immersive and true-to-life
sickness is an important consideration in cloud VR. The
experience. Allowing users to experience this anywhere
target motion-to-photon (MTP) latency, close to the limit
and anytime requires support from the 6G mobile system.
of human perception, is approximately 10 ms, which is half
New applications such as mobile 3D navigation will require
that of current VR requirements. In addition to requiring
3D images to be transmitted over mobile networks, giving
extreme video resolution and color depth, ultimate VR is
rise to extremely high requirements in terms of network
expected to require more than a 100-fold increase in the
bandwidth. The raw data rates, depending on image size,
raw data rate. Furthermore, an architecture that enables
resolution, color, and so on, will vary from sub-1 Tbit/s
pure remote rendering is more suitable for devices that
to a few hundred Tbit/s [14]. Research on compression
have limited computing capabilities — user devices often
techniques that can reduce the bandwidth consumption is
have strict constraints in terms of power and weight. In
ongoing.
this case, a stringent transmission latency (an RTT of less
than 2 ms) and higher data rate will be required.
• Broadband wireless access for the
• A¼ŅŔąÙ¼ģßġřěŔą̀ŌãģŌīƔ unconnected
ÙīġġřģąÙ¼Ŕąīģ Today, about 4 0% of people around the world lack
access to mobile networks. Ambitious plans to integrate
Haptic communication involves the exchange of real-time
terrestrial and non-terrestrial networks in the 6G era aim
haptic information, including surface, touch, actuation,
to increase the coverage rate to nearly 100% worldwide,
motion, vibration, and force information. This information
even in sparsely populated areas, expanding financial and
is transmitted over the network along with audiovisual
social inclusion. For people in remote unconnected areas
information. For example, haptic clothing can make a
or ships, non-terrestrial networks can serve as relay links
virtual football game feel more realistic, enabling the
to terrestrial base stations.
wearer to experience the texture, weight, and pressure
of the virtual ball, or it can allow the wearer to receive a Directly connecting non-terrestrial networks and mobile
virtual hug from a family member far away. phones is an attractive prospect, as it ensures seamless
s witchover be t we en dif ferent acce s s s er vice s . For
Among haptic applications, teleoperation with interactive
instance, it is important for people on the move to have a
feedback (such as tele-surgery, tele-diagnosis, and tele-
broadband connection. The integrated 6G system should
motion-control) in highly dynamic environments is the
provide optimal, scenario-specific MBB coverage for
most challenging. In these cases, haptic feedback is
people in cars, trains, planes, and ships.
important to stimulate the human brain and help users

11 | Huawei Technologies
Technology
Furthermore, with the integration of terrestrial and non- to wireless — for example, 6G. This requires ultra-high
terrestrial networks, 6G will be resilient against natural reliability (e.g., greater than 99.9999%) and low-latency
disasters, ensuring continuity of services. (e.g., sub-ms or even µs) deterministic communication
capabilities so that precise and reliable control can be
achieved.
Usage Scenario 2: URLLC+
6G will accelerate the comprehensive digital transformation • Collaborative robots in groups
of vertical industries. This usage scenario is the continuous
evolution of ultra-reliable low-latency communications In the fac tor y of the future, most of the major work
(URLLC) for critical machine-type communication (MTC) will be performed by robots instead of humans. During
in Industry 4.0 and beyond [15]. It also applies to new p ro d u c t i o n , n u m e ro u s t y p e s of ro b ot s — s u c h a s
applications enabled by the ubiquitous utilization of automated guided vehicles (AGVs) and drones — will
robots, UAVs, and new human-machine interfaces (HMIs) transport raw materials, spares, and accessories from the
in manufacturing, public service, autonomous driving, warehouse to the production line. For large or heavy parts,
and household management. To be more closely adapted multiple robots will collaborate to transport them — this
to all kinds of vertical applications, the requirements on is known as collaborative carrying [16]. To achieve safe
low latency and high reliability may be strict in first-order and efficient cooperation among these robots, a cyber-
statistics (e.g., mean number of errors in a period) but physical control application will be used to control and
controllable in the distribution or higher-order statistics coordinate their movement. For example, carrying rigid
(e.g., distribution of errors in a period). or fragile parts requires precise coordination, whereas
flexible or elastic parts allow a certain degree of freedom
for higher efficiency. To maintain the level of accuracy
• ;¼ÙŔīƔīøŔāãøřŔřňã needed for complex collaborative work, it will be necessary
to leverage the synchronization, latency, and localization
Unlike traditional assembly lines, which are suited to mass
accuracy capabilities provided by the 6G network. In this
production, factory of the future aims to implement full
case, a localization accuracy of 1 cm, an E2E latency of
automation and flexibility, meeting the demands of mass
approximately 1 ms, and reliability greater than 99.9999%
customization. To enable this revolution, the 6G network
may be desirable.
will play a key role. The precondition for modules to freely
move around in order to instantly form a customized
assembly line is the use of ultra-high performance radio • ;ňīġąģŔãěěąùãģŔÙīØīŔŌŔīÙŸØīňùŌ
links, which untether machines from interconnection
cables. Fur thermore, with AI and digital twins, it will Recently, collaborative robots — known as cobots —
be possible to accumulate and share manufac turing have appeared in the manufacturing industr y. Unlike
experience and knowledge among machines and robots, traditional robots that work in separate and restricted
helping optimize the evolving manufacturing process. 6G regions, cobots can collaborate and interact with people
could also bring many other benefits to the factory of the in close proximity. Like co-workers, they are expected to
future. For example, a ubiquitous RF sensing system would be intelligent (allowing them to understand the dynamic
enable proactive maintenance of the entire production environment and tasks), c autious to human safet y,
environment and processes. And, as the factory of the proactive to actions and risks, and reliable in functionality.
future requires no human onsite, lights-out manufacturing To achieve all of this, it is necessary to integrate AI, ICT,
would significantly lower the OPEX and carbon footprint. and OT. Furthermore, the high-performance sensing and
communication technologies integral to 6G are essential to
support cobots' mobility and interaction with humans.
• Motion control
Cyborgs, a concept laid out in 1960 [17], are the next
In addition to being one of the most challenging use evolutionary step of cobots. They are cybernetic organisms
cases, motion control is the core logic in the automation — humans enhanced with machines. For example, cyborgs
field [16]. It is responsible for controlling every aspect of a could be used to enhance a person's strength or sensory
machine's movements in a well-defined manner. This type abilities, or help someone overcome physical disabilities.
of operation already exists in modern manufacturing, but With the development of neuroscience, 6G will be the key
it is implemented via wired technologies such as industrial for cyborg interconnection.
Ethernet. In order to realize a truly flexible production
line, communication needs to be transformed from wired

Huawei
HuaweiTechnologies
Technology | 12
 • Vãűãěɧ¼řŔīģīġīřŌűãāąÙěãŌ manufacturing, and agriculture. The required data rate
could range from very low to medium, and the packet
In terms of technical requirements, autonomous driving arrival time inter val could range from a day to a few
is the most challenging use case of smart transportation. milliseconds. A key requirement is for sensors to have a
The initial level of autonomous vehicles (typically used long lifetime, but this may differ significantly depending
in scenarios such as mining, quarr ying, construction, on their energy harvesting capabilities. In some cases,
and agriculture) requires remote human driving and zero-power backscattering-based passive IoT devices
teleoperation. would also be applied as an option for extremely low-cost
connections.
Level 5 autonomous vehicles are a more advanced use
case, which will redefine the meaning of traveling by car.
• …ġ¼ƑØřąěßąģùŌ
As autonomous vehicles completely take over driving and
route planning, journeys in such vehicles could be relaxing,
Smar t building refers to managing and controlling a
enjoyable, and productive, while retaining the advantage
building as intelligent entities with seamless information
of a private space. To deal with unforeseen situations,
f lowing among related par ties, including elec tronic
sensing and AI capabilities provided by 6G, as well as
products, smart materials, control systems, and users.
ultra-low latency, high reliability, and precise localization,
Integration is the first step in making a building smart. As
will be essential.
a complex ecosystem, one building might contain many
different subsystems, including surveillance cameras,
‘Ō¼ùã…Ùãģ¼ňąīɥ˜ġ\Œ͚ elevator control, air conditioning, and electrical power.
The usage of 6G in the smart building industry should
6G will continue the journey started by 5G to connect enable a common infrastructure with high efficiency and
ever y thing, but it will do so with a broader variet y intelligence to be built. In addition, due to the massive
of devices, new HMIs, higher density of connections, number of sensors installed in a smart building, they will
and native tr us t wor thines s . This usage scenario is need to support large-scale connectivity and low energy
the continuous evolution of massive machine type of consumption. The second step is to interconnect buildings.
communication (mMTC), which is characterized by the In the future, mobile communication infras truc ture
massive number of lightly connected devices with sporadic will provide the digital foundation for cross-platform
traffic in smart cities, healthcare, buildings, transportation, trustworthiness.

13 | Huawei Technologies
Technology
• …ġ¼Ƒāã¼ěŔāÙ¼ňã ser vices. And with autonomous driving capabilities,
massive UAVs are expected to be utilized in future logistics
Pervasive and customized healthcare services, free of to deliver packages over long distances. It is not beyond
geographical constraints, is the vision for smart healthcare the realm of possibility that such UAVs could land on top of
in the future. As the mobile communications system cars or buses to recharge during a long-distance delivery.
develops, it will enable various new use cases to emerge,
including dynamic monitoring of personal health, tele- • ªąßã̀ň¼ģùąģùDīŒŌãƒąÙãŌ
diagnosis and pathology inference, holographic medical
and recovery training, and tele-surgery. In particular, Another area that will significantly benefit from 6G's
with the new sensing and AI capabilities in 6G, real-time global seamless coverage is wide-range IoT services. For
analysis on patient data could prove extremely beneficial. example, 6G could enable the collection of information
Furthermore, use cases such as tele-diagnosis and tele- from buoys in the oceans to report container status during
surgery will significantly reduce the pressure in an aging ocean transportation or from sensors in forests or deserts
society, especially in regions that lack sufficient medical to forecast and prevent natural disasters in a timely
resources. manner. Wide-ranging IoT services will be extended to
such unconnected locations to better protect the world.
• …ġ¼ƑŌãƒąÙãŌãģ¼ØěãßØŸ‘©Ō
Usage Scenario 4: Sensing
UAVs, commonly known as drones, come in a wide variety
of sizes and weights, and they can be used in various Networked sensing creates a new type of usage scenario
sectors [18]. UAV applications may cover many fields, beyond communication. It covers a range of use cases
such as unmanned inspection for mining and exploration, such as localization for device-based or even device-
and aerial filming for media and entertainment. However, free targets, imaging, environment reconstruction and
the more advanced communication, sensing, and AI monitoring, and gesture and activity recognition [19]. The
capabilities delivered by 6G will see UAV applications sensing usage scenario adds new performance dimensions
evolving to take more responsibilities in our daily lives. For to the International Mobile Telecommunications (IMT),
instance, UAVs in 6G can act as mobile base stations and such as detection probability, and sensing resolution and
provide on-demand, high-capacity coverage to deliver live accuracy (in terms of range, velocity, and angles). The
streaming of XR services and high-accuracy positioning requirements of these dimensions vary from application

Huawei
HuaweiTechnologies
Technology | 14
 to applic ation. For loc alization and recons tr uc tion robots in an automated factory to easily retrieve parts on
applications in the future, high sensing accuracy and a warehouse shelf and install them accordingly [20].
resolution are required, whereas for imaging applications,
ultra-high resolution is the key. And for gesture and In addition to high-accurac y absolute loc alization,
activity recognition, detection probability becomes the applications such as automatic docking and multi-robot
top priority. cooperation also pose high requirements on relative
localization. When a swarm of robots collaboratively lift
and carry a complex-shaped mechanical part or a drone
• High-accuracy localization and docks with a moving vehicle that has a small landing
tracking margin, it is critical for each robot or drone to determine
its locations with respect to others.
Empowered with sensing capabilities, the 6G network
will be able to provide positioning services for device- Fur ther empowered by AI, future systems could also
based targets (similar to 5G) and localization services provide semantic localization with context awareness and
for device-free objec ts (similar to radar use cases). dynamic address resolution according to service context.
Latency, Doppler, and angular spec trum information This would enable robots in restaurants to function similar
from scattered and reflected wireless signals can be to human waiters. For instance, a robot could deliver a
processed to extract coordinates, orientation velocity, glass of wine to a customer sitting by the window without
and other geometric information in a physical 3D space. requiring coordinates from a human.
High-accuracy 3D localization and tracking down to the
centimeter level enables meaningful association between
• …ąġřěŔ¼ģãīřŌąġ¼ùąģù˛ġ¼ŅŅąģù˛¼ģß
cyber information and the locations of physical entities. As
such, this will make various applications feasible, spanning localization
from factories to warehouses, hospitals to retail shops,
and agriculture to mining. For example, this could enable With simultaneous imaging, mapping, and localization,
three sensing capabilities are mutually enhanced. The

15 | Huawei Technologies
Technology
 human eye. With the THz ISAC technology that leverages
the mmWave band, such capability could be integrated
into a portable or wearable device, or even used in an
implant.

The "seeing beyond eyes" concept — enabled by ultra-

high resolution imaging — can be applied in daily life
to f ind pinpr ick leaks in water pipes or implement
contactless flaw detection and quality control in smart
factories. This concept also means making the invisible
visible. By utilizing the penetration characteristics of
electromagnetic waves, applications such as security
scans on packages or detection of cables in walls could be
performed on portable devices or even smartphones. The
larger radio frequency range being explored in 6G opens
up the possibility to see through materials such as skin,
subcutaneous fat, suitcases, and furniture. Spectrogram
recognition is another part of this concept and is based
on identifying targets through the spectrogram sensing
of their electromagnetic or photonic characteristics. The
unique absorption characteristics of different materials
can be characterized by THz signals. Typical applications
include food calorie detection and environmental PM2.5
analysis [21].

One of our prototypes, showcased at EuCNC2021 virtual

booth (virtual.eucnc.eu), has demonstrated the feasibility
of implementing THz imaging on por table devices .
imaging function captures images of the surrounding Leveraging communication waveforms, this prototype can
environment, and the loc alization func tion obtains sense and image a hidden object with mm-level resolution.
locations of surrounding objects. The mapping function More details can be found in [22].
then uses these images and locations to construct a map,
which in turn helps the localization function improve the • Gesture and activity recognition
inference of locations.
Device-free gesture and activity recognition using ML is
S imultane ous loc alizat ion and mapping (S L AM)
the key to promoting next-generation human-computer
applications in the mmWave or THz bands enable sensing
interfaces. It allows users to conveniently interact with
devices to construct 3D maps of the surroundings in
devices through gestures and actions. Such recognition
unknown environments. In the context of 6G, sensing
is divided into t wo t ypes: macro and micro. Macro
devices could be 6G base stations or terminals, including
recognition refers to body movements. One example of
cars, drones, and robots. Compared with the traditional
this is to automatically supervise patient security in future
lidar and optical camera systems, SLAM via 6G wireless
smart hospitals — for example, to detect falls or monitor
signals enables autonomous vehicles to obtain ultra-
rehabilitation exercises. Compared with traditional camera-
high resolution and accuracy in all weather conditions
based monitoring, one of the key benefits of ISAC is privacy
and the capability to see what is around the corner.
protection. Micro recognition refers to gestures, finger
Similar func tions for indoor scenarios would allow
movements, and facial expressions. Imagine playing pianos
robots and AGVs to move around freely even in crowded
and drawing pictures in the air but hearing real music and
environments.
seeing real pictures created by XR at the same time. This
will make it possible to truly create art anywhere anytime.
• řùġãģŔãßāřġ¼ģŌãģŌãŌ

Augmented human senses aim to provide a safe, precise,

and low-power sensing capability superior to that of the

Huawei
HuaweiTechnologies
Technology | 16
 Usage Scenario 5: AI network (DNN) models for each application. Following this
trend, in 10 years, AI model providers will be field-specific
T h i s u s a g e s c e n a r i o a i m s to i n te l l i g e n t l y c o n n e c t OTTs or carriers, while AI inference capabilities will be
distributed intelligent agents in order to proliferate large- delivered to individuals and vertical industries by public
scale deployment of AI in all industries. Spectrally efficient, service carriers or operators. A mobile communications
high-capacity, and low-latency transmission for distributed system that provides AIaaS for distributed learning and
learning — including data and model parameter exchange inference applications will be the key to meeting the real-
among large numbers of intelligent agents — is expected time and large-scale learning and inference requirements
for real-time AI. Native trustworthiness, with the support of society and vertical industries in the future. In terms
of native security and local data privacy, is a key enabler of distributed learning and inference services, the mobile
for this usage scenario. communications network is not simply a big pipe to
transmit bits and bytes; instead, it is a platform with
integrated connec tivit y and computing capabilities
• D̀ãģā¼ģÙãßģãŔŲīňė¼řŔīġ¼Ŕąīģ designed to provide optimal resource scheduling in
order to support learning tasks and achieve fast learning
Today, mobile networks require large workforces for convergence. The benefits of this will go beyond the
network operation, administration and maintenance superior performance (e.g., ultra-low latency) achieved by
(OA&M). AI has great potential to relieve this major labor bringing AI services closer to end users, while also meeting
and financial burden. For instance, the network system local privacy protection requirements.
itself could implement, operate, and manage network
configurations and function deployment. Manual passive
OA&M will evolve into zero-touch proactive OA&M — for ɦ˚^ãŲ)ěãġãģŔŌ
example, by using predictive network analytic services
and E2E system OA&M across all technical domains. AI in To meet the challenging requirements discussed earlier,
6G will adapt to environmental changes and optimize both in a d di t ion to t h e deve lop m e nt s ma de in w ire le s s
the communication and computing resources for optimal transmission technologies, the 6G system will encompass
solutions that meet diversified requirements. many new elements such as new spectrum, new channels,
new materials, and new devices. In this section, we discuss
the candidates and challenges in each of these aspects.
• D¼¼…øīňß¼Ŕ¼ġ¼ģ¼ùãġãģŔ

Huge volumes of data will be generated, collected, and ^ãŲ…ŅãÙŔňřġ

exchanged in future mobile communications networks.
Such data will be used to perform and optimize various As the primar y resource, spec trum is the main
network services related to operation and management consideration for each generation of wireless
t a s k s (e . g . , f o r c o n f i g u r a t i o n m a n a g e m e n t , f a u l t communication technology — more spectrum is needed
management, and SLA assurance). Moreover, knowledge to support higher data rates. Furthermore, a globally
ex trac te d from raw data could be exchange d with unified process for spectrum allocation is critical for
other systems or business sectors in order to generate greater economies of scale and more convenient global
a broader scope of value. Data is an essential asset for roaming, both of which are important factors in realizing
AI, but not all the raw data is high quality or usable. As a succe s s ful indus tr y on a global sc ale. A s mobile
such, it is necessary to support efficient data processing communication technologies evolve to new generations,
while reducing computational complexity and energy the use of spec trum continues to expand to higher
consumption. In this regard, AIaaS could be used to select frequency bands. While 5G was the first to use mmWave
high-quality data from vast amounts of raw data. frequency bands, 6G is expected to explore even higher
frequency bands, such as (sub-)THz. Although low-
and mid-band frequencies are impor tant for mobile
• AIaaS for distributed learning and
communication systems to achieve wide coverage, 6G
inference will apply the concept of a multilayered frequency band
framework, as illustrated in Figure 12 and described later.
In the 6G era, software generation will be transformed
from Software 1.0 (human coding) to Software 2.0 (data • VīŲ̀¼ģßġąß̀ؼģßŌňãġ¼ąģŔāãġīŌŔÙīŌŔ̀ãøøãÙŔąűã
coding), in which massive data is provided to deep way for wide coverage. Low bands (from 700 to 900
learning algorithms in order to generate deep neural MHz) and mid-bands (from 3 to 5 GHz) play a crucial

17 | Huawei Technologies
Technology
 Figure 12 Multilayered frequency band framework

role in 5G and are expected to be vital in 6G as well. allocated to mobile services in the THz range of 100 to
Toward 2030 and beyond, at least 1 to 1.5 GHz of 450 GHz. This makes it possible to support very high
additional mid-band spectrum is needed to support data rates for short-distance (less than 10 meters)
the continued grow th of traffic, especially when and mid-distance (e.g., 200 meters) communication.
considering multi-operator coexistence. The 6 GHz In addition, THz bands bring enhanced sensing
(i.e., 5925–7125 MHz) and 10 GHz (i.e., 10–13.25 GHz) resolution thanks to ultra-wide bandwidth and shorter
bands are competitive candidates. Compared with wavelengths. In the future, smart phones integrated
3.5 GHz, propagation attenuation will be increased with THz sensing technology will be able to augment
in an acceptable range while path loss will be further human senses — for example, they will detect calories
reduced by more advanced radio technologies. in food, find pinprick leaks in water pipes, facilitate
security checks, or monitor the skin and subcutaneous
• ġġª¼űãØãÙīġãŌġ¼ŔřňãąģŔāãɨ<ãň¼˚ Compared vascular health.
with low- and mid-bands, the mmWave band is more
challenging due to more severe radio propagation
characteristics. However, new drivers will emerge New Materials and Antennas
in the 6G era. First, a significant volume of available
bandwidth in the mmWave bands is essential for The tremendous evolution of digital communication over
the ultra-high data rates required in 6G. Second, the past few years can be attributed to the remarkable
mmWave bands are the key spec tr um that c an progress made in semiconductor technologies. With 6G
achieve a centimeter-level sensing resolution, which on the horizon, new material technologies will continue to
is especially important for mapping with network evolve, facilitating the application of new spectrum and
infrastructures. This is difficult for mid-bands due to new antennas for new usage scenarios.
practical limitations such as available bandwidth and
antenna aperture size. Third, the evolution of more • …ąěąÙīģ¼ßű¼ģÙãġãģŔŔīŲ¼ňߌAƀ
advanced radio techniques can also improve the
utilization of the mmWave bands. E-bands (71–76 and Silicon technologies, which inherently have low cost, high
81–86 GHz) are prime candidates to support larger yield, small geometry, and low power, have been used
contiguous blocks in the future, where integrated to continuously drive next-generation applications in
access and backhaul (IAB) would be a key technology communication, imaging, computing, and more. Based
for efficient spectrum utilization. on the existing silicon platform — one that is already
mature — advanced proces s features are added to
• THz bands open new possibilities for sensing and
enable new capabilities. For example, unlike the standard
ÙīġġřģąÙ¼Ŕąīģ˚ One of the most notable features of
complementary metal-oxide-semiconductor (CMOS), the
the THz bands is the potential to provide ultra-wide
SiGe-BiCMOS platform can now successfully perform many
bandwidth. About 230 GHz of spectrum has been
applications simultaneously, such as imaging, spectroscopy,

Huawei
HuaweiTechnologies
Technology | 18
 Figure 13‘ŌãÙ¼Ōããŷ¼ġŅěãīøD…Ōãģ¼ØěãßØŸňãÙīģƍùřň¼Øěãġ¼Ŕãňą¼ěŌ

and communication. Advanced processes allow for a more surface wave will interfere with the antenna radiation and
efficient and compact hybrid integration of both photonic result in poor performance. Antenna-in-package is another
and electronic components on the same silicon, something possibility, although the interconnection loss between
that is predicted to be realized in the near future. antennas and monolithic microwave integrated circuits
(MMICs) is high. At THz frequencies, it is both desirable
Silicon technology, however, has fundamental limitations and challenging to realize the design and implementation
for photonic use due to its indirect bandgap. Type III– of efficient and low-loss THz antennas.
V semiconductors such as InP and GaAs with a direct
bandgap were therefore proposed to overcome
these limitations. But the high cost involved in such
• ãÙīģƍùřň¼Øěãġ¼Ŕãňą¼ěŌ¼ģß
semiconductors has prevented their wide adoption across ąģŔãěěąùãģŔŌřƐ¼ÙãŌ
the market. To overcome the limitations of silicon while
leveraging photonic features, heterogeneous integration of Tuning of materials' electrical properties is desired in many
silicon with III–V semiconductors has emerged, combining cases because it allows for devices with more functions,
the advantages of both. Integrating III–V materials on the smaller dimensions, and reduced costs. As such, various
same silicon wafer in a standard lithography process has tunable materials have been proposed and embedded
shown great potential in many photonic applications. into systems for flexible and dynamic control. This tuning
feature has enabled reconfigurable intelligent surfaces
The advancements of semiconductor technologies have (RISs), which are controlled through a digital platform.
made it possible for us to achieve a THz integrated circuit R I S s c an manipulate t he inc ide nt e le c t romagne t ic
(IC). We can now fabricate ICs up to 700 GHz through waves to desired outputs through carefully designed
SiGe heterojunction bipolar transistor (HBT) technology. elec tromagnetic scatterers (meta-atoms), which are
Estimates indicate that the performance limit of SiGe HBT designed to induce phase or amplitude changes (or
may reach or even exceed 1 THz in the near future [23]. both) on the incident waves and can therefore perform
Silicon THz ICs hold several advantages, such as low beamforming and steering. These scatterers can be made
cost, compact size, high yield, and easy integration. A of tunable materials and controlled elec tronically or
convenient way to implement a THz antenna involves thermally. For example, graphene, liquid crystal, and phase
directly integrating it with the frontend circuit on a silicon change materials have been used for such surfaces and
substrate. However, on-chip antenna design is challenging demonstrated feasible for realizing dynamic control.
due to the surface wave generated in substrates. The

19 | Huawei Technologies
Technology
 Figure 14 New challenges in channel modeling for 6G

Figure 13 shows a use case example of RISs enabled ray tracing, and measurements. Stochastic models are
by reconfigurable materials [24]. The RISs can be used usually built upon the distribution of scattered clusters,
to extend the coverage from outdoor or indoor base which are randomly generated by a specified probability
stations to users, vehicles, or AGVs in cases where there density function. Hybrid models (also known as quasi-
is no direct link between them or the link is blocked by deterministic models) are a combination of the other two
obstacles. By tuning the phases of elements, the RISs can models — they typically combine the dominant paths
direct their beams to the target end users dynamically calculated by deterministic models and scattering paths
and relay information to the desired locations with generated by stochastic models.
at tenuation compensation. Beamforming at a base
station and reflecting phase control at the RISs can be In terms of large-scale system-level evaluation, stochastic
jointly optimized to maximize multi-user performance. In models, as adopted in 3GPP, are usually simpler and more
addition, the potential large apertures may help enhance efficient than deterministic models. However, these models
resolutions in sensing applications. cannot express the deterministic parameters related to
a specific system or scenario. For example, they cannot
express geometrical information related to multipath
New Channels channel parameters or locations of communication devices
or scatterers. Deterministic models are usually applied
Radio wave propagation is a fundamental part of wireless when precise characterization of the channel environment
communications. Before constructing and operating real- is needed, but this increases computational complexity.
world systems, we must understand the principles of radio From 3G to 5G, research on the channel model tended to
propagation and develop the associated channel models. focus on improving the deterministic levels under limited
These models represent the key propagation processes complexity.
and allow for meaningful evaluation of and comparison
between different systems [25]. In 6G, channel modeling faces new challenges introduced
by the potential new spectrum, antennas, and scenarios, as
The methodology of channel modeling is roughly divided illustrated in Figure 14. A single type of channel modeling
into three categories: deterministic, stochastic, and scheme may not be sufficient to meet the evaluation
hybrid. In deterministic models, the physical propagation requirements of all usage scenarios in 6G. In contrast,
parameters are fixed, meaning that the real physical usage scenario-dependent channel modeling might be a
channels in specific scenarios can be reconstructed using viable option, while a potential hybrid model that achieves
techniques such as computational electromagnetics (CEM), a good trade-off between accuracy and complexity is

Huawei
HuaweiTechnologies
Technology | 20
 Figure 15 Capabilities and trends for future devices

worthy of investigation. Some examples and challenges such as cross-polarization rates, incidence-angle-
are discussed below. dependent phase shifters, and non-ideal impairments
will be important features that need to be modeled
• ^ãŲ ŌŅãÙŔňřġ˜ As the spectrum goes beyond 100 in order to determine the beneficial application
GHz into the THz bands, the free space path loss scenarios of RISs.
increases accordingly. To compensate for this impact,
more advanced beamforming technologies would • New scenarios: New scenarios such as ISAC depend
be needed; other wise, the applied range will be heavily on the surrounding environment, especially
limited. One challenge in particular for the THz band for scatterer distribution, which is difficult to describe
is the so-called molecular absorption phenomenon. through stochastic models. In this case, deterministic
This is where THz signals excite gas molecules in the models related to some specific geographical areas
atmosphere, converting part of the signal power into are preferred. Which type of model to use for sensing-
kinetic energy of the gas molecules. In terms of small- assisted communication is still an open question. In
scale fading, measurement results [26][27] show that, addition, for NTN scenarios with satellites, the upper
similar to mmWave signals, THz signals also exhibit atmosphere and clutter loss in the propagation model
multipath propagation characteristics especially in need to be considered, while for drones serving as
indoor scenarios, enabling multi-stream transmission base stations, new channel models with moving base
in THz bands. stations would be developed.

• New antennas: With the development of ultra-

massive MIMO, new antenna s truc tures such as New Devices
extremely large aperture arrays (ELAAs), RISs, and
orbital angular momentum will significantly affect The next revolution of mobile devices, like smartphones
channel modeling and performance evaluation. For replacing feature phones in the past 20 years, is expected
instance, ELAAs bring new channel features, such to occur in the 6G era. It is anticipated that future devices
as near-field spherical waves and non-stationary will adopt new capabilities — such as sensing and imaging,
channels, which need to be characterized. For RISs, haptic communications, holographic display, and AI —
in addition to the near-field effects, scattered models that are enabled by the 6G communications system. These
new capabilities may also include human-level perception,

21 | Huawei Technologies
Technology
ambient sensing, multimodal human-machine interaction, functions, such as imaging, spectroscopy, healthcare
and energy har vesting, as shown in Figure 15. Such monitoring, and gesture recognition.
capabilities will transform today's devices, which function
as "intelligent assistants connecting physical and cyber • Diversified: This refers to not only smartphones,
worlds" into ones that function as "hyper terminals in a but also various types of devices that act as sensors
converged physical-cyber world". In the following section, and actuators. In the future, a wide range of human-
we discuss the four major trends that are driven by the centric and industrial devices will emerge, integrating
development of these new capabilities. advanced sensors, new display technologies, and
AI — for example, wearable devices, implantable
• …ġ¼ňŔãň˜This refers to not only making smartphones medical devices, automobiles, robots and cobots, and
smarter, but also augmenting realities to automate smart factory equipment. The explosive growth of
everything. In the future, mobile devices will be able to diversified devices will pose higher requirements on
implement AI capabilities and offload computationally interconnectivity. Anchor devices will help to provide
intensive tasks to edge clouds. With AI/ML and a seamless and consistent user experience.
the development of shor t-range communication
technologies in 6G, devices in the future will have • Cloudified: This refers to not only physical devices,
greater intelligence. They will also automate more but also virtual devices that enable privacy protection
aspects of our life, improving our service experience and new business models. In the future, each 6G
and productivity. device will have a virtual counterpart in the cloud
acting as its proxy. Such devices are shared in public
• Ve r s a t i l e : T h i s r e f e r s t o n o t o n l y p r o v i d i n g places and used on demand. With virtual devices in
c o n n e c t i v i t y, b u t a l s o o f f e r i n g n ove l s e n s i n g the cloud, users can access desired services anytime
capabilities to open up new possibilities for future and anywhere via shared devices.
mobile applications. In the future, devices with multi-
sensory capabilities could be integrated into humans In addition to devices, the interfaces between them and
to advance the human race, forming cybernetic humans will also evolve. The brain-computer interface
organisms. Novel sensing capabilities will create the
potential for mobile devices to support many new

Huawei
HuaweiTechnologies
Technology | 22
 Figure 16~¼ň¼ßąùġŌāąøŔŌīøøřŔřňã¼ąňąģŔãƐ¼ÙãßãŌąùģ

(BCI) concept, first introduced in the 1970s, has undergone enable tailored optimization of air interfaces for different
significant development. Today, with the help of better users. The personalized air interface can customize the
neural knowledge and novel neural sensors, implantable transmission scheme and parameters at the UE-level to
multimode sensory neurochips (integrating touch, smell, enhance experience without sacrificing system capacity.
sight, hearing, and taste functions) are not far from reality. Fur thermore, it can be scaled easily to suppor t the
This will enable the human brain to communicate directly near-zero-latency URLLC. In addition, a new signaling
with devices in the future, a quantum leap from today's mechanism — one that is both simple and agile — will
interaction via a screen and keyboard on smartphones. minimize the signaling overhead and delay.

ɧ˚)ģ¼ØěąģùŒãÙāģīěīùąãŌ¼ģß • ;ňīġ¼ßß̀īģDīŅŔąġąƀ¼ŔąīģŔī
Architectures native AI

For 6G, AI will be a built-in feature of air inter faces,

In this section, we focus on the paradigm shifts in air
enabling intelligent physical layer (PHY) and media access
interface and network architecture designs. More detailed
control (MAC). Rather than being limited to network
discussions on the enabling technologies and architectures
management optimization (such as load balancing and
are provided in [1].
power saving), it will also replace some non-linear or non-
convex algorithms in transceiver modules, or compensate
~¼ň¼ßąùġ…āąøŔŌąģŔāãąňDģŔãƐ¼Ùã for deficiencies in the non-linear models. AI will not
only make 6G PHY more powerful and efficient, it will
Design also facilitate the optimization of PHY building blocks'
parameters. Furthermore, it will help provide new sensing
To serve the diversified use cases discussed in Section 3
and positioning capabilities, which in turn will significantly
and meet the challenging goals toward an era featuring
change the design of air interface components. AI-assisted
connected intelligence, revolutionary breakthroughs are
sensing and positioning will also make it possible to
required in 6G air interface design, triggering paradigm
implement low-cost and highly accurate beamforming and
shifts in the design philosophy, as highlighted in Figure 16.
tracking. In addition, intelligent MAC will provide a smart
controller based on single- or multi-agent ML, including
• ;ňīġŌīøŔ¼ąňąģŔãƐ¼ÙãŌŔīąģŔãěěąùãģŔ cooperative ML for network and UE nodes. For example,
with multi-parameter joint optimization and individual- or
¼ąňąģŔãƐ¼ÙãŌ
joint-procedure training, enormous performance gains can

The 6G air interface design, powered by a combination

of model- and data-driven AI capabilities, is expected to

23 | Huawei Technologies
Technology
be obtained in terms of system capacity, UE experience, sensing, advanced positioning, and AI, we can replace
and power consumption. the conventional beam sweeping, beam failure detection,
and beam recovery approaches with proactive UE-centric
beam generation, tracking, and adjustment schemes. In
• ;ňīġ¼ßß̀īģŅīŲãňŌ¼űąģùŔīØřąěŔ̀ąģ addition, "handover-free" mobility can be realized at the
power saving physical layer, at the very least. New intelligent UE-centric
beamforming and beam management technologies will
Minimizing power consumption for both network nodes enhance UE experience and overall system performance.
and terminal devices should be a key requirement in Moreover, the emerging RIS and moving nodes such as
the design of 6G air interfaces. Unlike the power saving drones make it possible for us to shift from passively
mechanism used in 5G, where power saving is an add- dealing with channel conditions to actively controlling
on feature or optional mode, power saving in 6G will them. In this case, the radio transmission environment can
b e a buil t-in feature an d de faul t op e rat ion mode. be changed to create the desired transmission channel
With intelligent management of power utilization, an conditions for optimal performance.
on-demand power consumption strategy, and other
new enabling technologies (such as the sensing- and
positioning-assis ted channel sounding scheme), we
• ;ňīġňã¼ÙŔąűãÙā¼ģģãěŔň¼ÙėąģùŔī
anticipate that both the network and terminals in 6G will active channel sensing/prediction or
feature significantly improved power utilization efficiency.
controlling

• ;ňīġÙīġġřģąÙ¼ŔąīģīģěŸ In order to achieve highly reliable wireless communications,

it is imperative to have accurate channel information. But,
to integrated sensing and
due to the delay involved in measurement and reporting,
ÙīġġřģąÙ¼Ŕąīģ and the pilot overhead (especially for high-speed mobile
UEs), obtaining real-time channel information is difficult.
The communication network as a whole can serve as a Sensing- and positioning-assisted channel sounding
sensor with high resolution and good coverage. It can powe re d by A I c an t ran s for m pilot-ba s e d c hann e l
be viewed as a sensing network that generates useful acquisition into location-aware channel acquisition.
information (such as locations, Doppler, beam directions, With the information obtained from sensing/positioning,
and images) for new services such as localization, tracking, we can dramatically simplify the beam search process.
and environment monitoring while also functioning to Furthermore, proactive channel tracking and prediction
assist communications. In addition, the sensing-based can provide real-time channel information and eliminate
imaging capability on terminal devices offers new device channel aging. In addition, new channel acquisition
functions such as imaging. A new design requirement technology minimizes both channel acquisition overhead
for 6G involves building a single network that integrates and power consumption for network nodes and terminal
both sensing and communication functions under the devices.
same air interface design framework. We hope that a
carefully designed communication and sensing network
will offer full sensing capabilities, while also meeting all
• ;ňīġÙãěěřě¼ňŅěřŌŌ¼ŔãěěąŔãŌŸŌŔãġŌ
communication KPIs more effectively. to integrated terrestrial and non-
ŔãňňãŌŔňą¼ěŌŸŌŔãġŌ
• ;ňīġŅ¼ŌŌąűãØã¼ġġ¼ģ¼ùãġãģŔŔī
Recent 5G standard releases have been ex tended to
‘)̀ÙãģŔňąÙØã¼ġīŅãň¼ŔąīģŌŲąŔā
include satellite networks but as separate systems [28]. As
controlled channel a step further, by integrating terrestrial and non-terrestrial
systems, it is expected that 6G will achieve universal
Beam-based transmission is impor tant for high coverage and on-demand capacity. In 6G, the satellite
frequencies such as mmWave. Major efforts are needed in constellation will be viewed as a new type of network
order to generate and maintain the precise alignment of node due to the tight integration of terrestrial and non-
the transmitter and receiver beams. Beam management terrestrial systems. Combining the designs of both systems
in 6G, however, is likely to be more challenging due will make multi-connection joint operations more efficient,
to the exploration of even higher frequency ranges. functionality sharing more flexible, and cross-connection
Fortunately, with the help of new technologies such as switching faster. This will go a long way in helping 6G

Huawei
HuaweiTechnologies
Technology | 24
 achieve global coverage and seamless global mobility with to account for RF impairments and restrictions. And,
low power consumption. given 6G's native AI capability, joint RF and baseband
optimization and design may be possible.

• ;ňīġġřěŔą̀Ù¼ňňąãňīŅãň¼ŔąīģŌŔī
ŌřŅãň̀ƏãŷąØěãŌŅãÙŔňřġřŔąěąƀ¼Ŕąīģ ~¼ň¼ßąùġ…āąøŔŌąģŔāã^ãŔŲīňė
Architecture Design
Intelligent spectrum utilization and channel resource
management are important design aspects in 6G. More I n a d d i t i o n to of fe r i n g t h e c o nve n t i o n a l ra n g e of
frequency bands, as mentioned in Section 4.1, will be connectivity services, 6G systems could also serve as
explored to provide larger bandwidths, which will support distributed platforms for executing user workloads in
the unprecedented data rates required by 6G. However, all industry scenarios. This is possible because the 6G
higher frequencies suffer from greater path loss and network will be built based on a decentralized and user-
atmospheric absorption. As such, we must consider how centric architecture that integrates native AI capabilities.
to effectively utilize these new spectrums together with With new enabling technologies, 6G will shift traditional
lower frequency bands when designing 6G air interfaces. paradigms toward a novel architecture that meets new
Furthermore, even though full duplex has been promoted requirements and integrates new capabilities, as shown in
in 5G, it is eagerly anticipated to become more mature in Figure 17.
6G.

• ;ňīġÙěīřß̀ÙãģŔňąÙDŔīģ¼ŔąűãD
• ;ňīġ¼ģ¼ěīù̀¼ģ߁;̀řģ¼Ų¼ňãŔī
In today's networks, AI services are located in a central
analog- and RF-aware
cloud at the application layer. In the 6G era, however,
network architecture and AI will go hand in hand. Put
Baseband protocols and algorithms are usually designed
differently, native AI support will be one of the fundamental
without carefully considering the features of analog and
factors that drive innovation in the network architecture.
RF components. This is because it is difficult to model the
As such, deeply converged communication and computing
impairments and non-linearity of such components. In
resources with a fully distributed architecture will lead to
6G, the design of the baseband physical layer is expected
a transformation from cloud AI to network AI. The benefits
of this will go beyond the superior performance (e.g.,
ultra-low latency) achieved by bringing AI services closer
to end users — privacy concerns could also be locally

Figure 17 Paradigm shifts of future network architecture design

25 | Huawei Technologies
Technology
resolved. This is one of the primary drivers of development security design. To guide the design of a trustworthy 6G
in terms of the 6G network architecture. However, the architecture and define the corresponding key capabilities,
architecture will be significantly impacted by privacy new use cases that yield new requirements, as well as new
and data governance requirements such as GDPR [4], enabling technologies, should be taken into account.
which advocates for the self-sovereign management of
personal data. This means that data ownership should be
returned to end users without any intervening authority.
• ;ňīġùãģãňąÙØąŔ̀ŅąŅãŔīřŌãň̀ÙãģŔňąÙ
Network AI holds especially true in terms of realizing real- ÙřŌŔīġąƀãßŌãƒąÙã
time AI functions, because training big data for ML and
executing AI inference are inefficient within the centralized From a functional perspective, the network manages the
cloud AI. state of each UE or end user. In this regard, the network is
essentially a large distributed state machine, meaning that
it maintains consistent states across different network
• ;ňīġąģøīňġ¼Ŕąīģ̀ÙãģŔňąÙÙīģģãÙŔąīģ functions. This requires the complex exchange of signaling
to task-centric connection messages, potentially limiting the extent to which network
per formance c an be improved (e.g., latenc y). More
Conventional communication systems, originally driven importantly, it may also lead to increased attack points
by voice and then data communication, mainly focus on (e.g., an increased attack surface). As the numbers of
information-centric connections. The communication connected devices and users increase, monolithic network
source and destination are clearly defined by end users functions (both physical and virtual) also become potential
and the services they intend to use or the other users with sources of serious bottlenecks. Because the network
whom they intend to communicate. As such, the entire inherently manages the state of each UE or user, we can
communication mechanism (e.g., session management and understand why a network design based on the per-UE or
mobility management) is designed to provide sufficient per-user perspective is needed. Specifically, a user-centric
support for this connectivity model. Conversely, the 6G design is capable of providing a virtual private network
system is expected to consist of numerous distributed (VPN) for each user, and such per-user VPNs deliver
nodes (e.g., terminals, radio access nodes, and network network services such as mobility management, policy
equipment) with intelligent features that provide native control, session control, and personal data management.
support for intelligent services or that utilize intelligence In addition, signaling overhead and the corresponding
for self-improvement. AI and sensing are two of the key network performance can be optimized at the per-user
services that 6G will provide. In order to provide these level.
services, the same task may be executed across numerous
dis tributed nodes in a coordinated manner. This is
referred to as task-oriented communication. In the future,
• ;ňīġīŅãň¼Ŕīň̀ÙãģŔňąÙűąãŲŔī
wireless communications technology should suppor t ŅňīŌřġãň̀ÙãģŔňąÙűąãŲ
diversified device types and time-varying topologies in
order to deliver optimal performance for task-oriented 6G will bring about a paradigm shift as it drives economic
communication. and social changes with advances in virtualization and
AI. 6G networks will have intelligence at their foundation,
e n a b l i n g a p a r t i c i p a to r y a p p ro a c h to n e t wo r k i n g
• ;ňīġŌãÙřňąŔŸ̀ÙãģŔňąÙ¼ňÙāąŔãÙŔřňãŔī and service provisioning. This will redefine the smart
ġřěŔąě¼Ŕãň¼ěŔňřŌŔ¼ňÙāąŔãÙŔřňã connec tivit y infras tr uc ture as a dynamic pool that
contains the resources of all participating users. It is a
5G security, implemented through a standalone framework, radical paradigm shift from the conventional operator-
is distinct from other network services. In 6G, one of the centric view to an inclusive prosumer-centric one, where
major paradigm shifts is the transition from simple security a "prosumer" is a combination of a "producer" and a
to native trustworthiness. This shift involves dealing with "consumer". Through a collaborative model bringing
the security-by-design framework and a wide range together many networks, key aspects such as multilateral
of topics, such as the trust model and security thread ownership, data ownership and privacy, and trust models
from the promising development of quantum computing of involved players must be designed as built-in features
and application of new technologies (e.g., AI and ML) in rather than add-on ones. Furthermore, in order to achieve

Huawei
HuaweiTechnologies
Technology | 26
 Figure 18…řġġ¼ƔīøŅ¼ň¼ßąùġŌāąøŔŌąģɨ<

local data governance and network sovereignty, 6G will trust architecture and post-quantum cryptography.
adopt new trust and security technologies. In an inclusive • Algorithms in each layer of the communication system
prosumer-centric model, every system participant can will change from analytic only to simultaneous model-
both contribute and consume resources and services. and data–driven ones, leveraging AI and ML to couple
Moreover, AI and ML technologies will enable autonomous with prac tical conditions that are hard to model
OA&M of 6G networks, involving little to no manual analytically.
intervention and allowing such networks to flexibly adapt
• Level of automation in network OA&M will be further
to everyone's needs. In this regard, 6G networks will be
upgraded toward fully touchless "level 5" native
tailored rather than proprietary, giving rise to the concept
automation.
of "my network".
• To nat i ve l y s uppor t inte llige nce in t he s ys te m
and provide AI as a ser vice for third par ties, the
ɨ˚…řġġ¼Ɣ¼ģ߁ī¼ßġ¼Ņ networking infrastructure will become converged
networking and computing infrastructure.
As Guglielmo Marconi said in 1932, it is dangerous to put • With the construction of mega-LEO constellations, the
limits on wireless. Over the last four decades, the wireless networking infrastructure will extend from terrestrial
revolution has reshaped our lives. Now, the next horizon of only to integrated terrestrial and non-terrestrial.
innovation will drive new paradigm shifts. In the last part
of this article, we summarize the paradigm shifts and lay ɨ<ī¼ßġ¼Ņ
out the expected timeline of 6G standardization in ITU-R
and 3GPP. Open, collaborative, and patient research will be Since 2018, numerous initiatives have been launched for
the key to the success and long-lasting value of 6G. 6G research. Industry and academic circles in Europe,
China, Japan, South Korea, and the USA have been
engaged in identifying the typical application scenarios,
…řġġ¼Ɣīø~¼ň¼ßąùġ…āąøŔŌ key capabilities, and potential technologies for the next-
generation wireless network. As the leading international
Here, we summarize the paradigm shifts the following
organization, ITU-R is initiating a new cycle toward 2030
aspects, as shown in Figure 18:
and beyond. ITU-R Working Party 5D has started the
study of Future Technology Trend and VISION for next-
• Services in 6G will change from connectivity only to
generation IMT standards.
connectivity plus sensing and AI.
• Private net working will be suppor ted from the According to the current schedule, ITU-R will complete the
extension of public networking to native design from VISION study in mid-2023, before World Radio Congress
day one. 2023 (WRC23) commences. It will provide a framework
• Encr yption-based security will transform toward and overall objectives for 6G, including usage scenarios
technology-based trustworthiness with multilateral and key capability requirements. As further study in the
industry continues to fully analyze how these requirements

27 | Huawei Technologies
Technology
 Figure 19 Expected timeline for 6G standardization

will affect the design of wireless communication systems,

3GPP may initiate an overall study into 6G, possibly around
the end of 2025 (R20). At present, we expect the first References
version of 6G standardization to be released sometime
around 2030. [1] Tong, W and Zhu, P. (Eds.). (2021). "6G: The next
horizon: From connected people and things to
The mobile communications industry — from 1G to 5G
connected intelligence." Cambridge: Cambridge
— has become relatively mature, and its grow th has
University Press.
slowed significantly. Unified standards have become
more important than ever for achieving economies of
[2] ITU-R. "IMT Vision – Framework and overall
scale. As with 4G and 5G, ensuring that 6G is a success
objectives of the future development of IMT
requires industry, academic, and ecosystem players to
for 2020 and beyond." Recommendation ITU-R
work together. Huawei firmly believes in and is actively
contributing to the continuous global cooperation in the M.2083-0, Sept 2015.
development of 6G. The standardization of 6G worldwide
[3] 6GANA. "From cloud AI to network AI, a view from
will undoubtedly be the path to success for decades to
come. 6GANA." Whitepaper, 2021 [Online]. Available:
āŔŔŅ˜˫˫ŲŲŲ˚ɨù̀¼ģ¼˚Ùīġ˫řŅěī¼ß˫ƍěã˫ɤɢɤɣɢɨɣɫ˫ɨɥɩɧ
ɫɨɫɦɧɪɧɢɧɣɫɥɨɨɨɪɧɣɧɤɩ˚Ņßø˚

[4] On the protection of natural persons with regard

to the processing of personal data and on the free
movement of such data, and repealing directive
95/46/ec (General Data Protection Regulation),
ãùřě¼Ŕąīģ˺ãř˻ɤɢɣɨ˫ɨɩɫ˛Ņňąěɤɢɣɨ˚

[5] )˚˚<ňąƆīň˛˚<ňããň˛%˚˚ªīěěġ¼ģ˛¼ģß\˚ē˚řňģŌ˚
"Framework for cyber-physical systems: volume

Huawei
HuaweiTechnologies
Technology | 28
 ɣ˛īűãƒąãŲ˛űãňŌąīģɣ˚ɢ˚̝^D…Œ…ŅãÙą¼ě~řØěąÙ¼Ŕąīģ˛ "Challenges and road ahead for wireless networks
ɤɢɣɩ˚ ŔīŌãƒãąġġãňŌąűãāřġ¼ģÙãģŔňąÙ¼ŅŅěąÙ¼ŔąīģŌ˚̝
2020 IEEE 92nd Vehicular Technology Conference
[6] NGMN. "6G drivers and vision." White paper 1.0,
(VTC2020-Fall), 2020, pp. 1-5, doi: 10.1109/VTC2020-
Apr. 2021 [Online]. Available: https://www.ngmn.
;¼ěěɦɫɩɤɪ˚ɤɢɤɢ˚ɫɥɦɪɪɧɢ˚
org/wp-content/uploads/NGMN-6G-Drivers-and-
©ąŌąīģ̀©ɣ˚ɢ̇ƍģ¼ě˚Ņßø˚ [15] 5G-ACIA. "Position paper: Our view on the evolution of
ɧ<ŔīŲ¼ňßŌɨ<˚̝;ň¼ģėøřƑ˜ɧ<̀D˛ɤɢɤɣ˚
˾ɩ˿ "Transforming our world: The 2030 Agenda for
…řŌŔ¼ąģ¼Øěã%ãűãěīŅġãģŔ˺˫)…˫ɩɢ˫ɣ˻˚̝‘ģąŔãß [16] ɥ<~~˚̝ɥ<~~Œ…ɤɤ˚ɣɢɦ̀…ãƒąÙãňãŇřąňãġãģŔŌøīň
Nations, October 2015. https://www.un.org/ga/ ÙŸØãň̀ŅāŸŌąÙ¼ěÙīģŔňīě¼ŅŅěąÙ¼ŔąīģŌąģűãƑąÙ¼ě
Ōã¼ňÙā˫űąãŲ̇ßīÙ˚¼ŌŅˡŌŸġØīě͞˫)…˫ɩɢ˫ɣΟV¼ģù͞) domains."

[8] T˚˚īīŅãň˛˚R˚%ãģùěãň˛^˚VěīġؼƑ˛˚Œāīġ¼Ō˛ ˾ɣɩ˿ M. E. Clynes. "Cyborgs and space." Astronautics, vol.

G. Chattopadhyay, and P. H. Siegel. "THz imaging ɣɦ˛ŅŅ˚ɩɦ̀ɩɧ˛ɣɫɨɢ˚
ň¼ß¼ňøīňŌŔ¼ģßīƆŅãňŌīģģãěŌÙňããģąģù˚̝DģD)))
[18] )˚\¼ƀ¼ňã¼ģř˚̝ąňŔň¼ģŌŅīƑ¼Ŕąīģ̂…Ŕ¼ŔąŌŔąÙŌΟø¼ÙŔŌ˚̝
Œň¼ģŌ¼ÙŔąīģŌīģŒãň¼āãƑƀ…ÙąãģÙã¼ģߌãÙāģīěīùŸ˛
2020. [Online]. Available: https://www.statista.com/
vol. 1, no. 1, pp. 169-182, Sept. 2011, doi: 10.1109/
ŔīŅąÙŌ˫ɣɩɢɩ˫¼ąň̀Ŕň¼ģŌŅīƑ¼Ŕąīģ˚
TTHZ.2011.2159556.

[19] D. K. Pin Tan et al. "Integrated sensing and

[9] H. Song and T. Nagatsuma. "Present and future of
communication in 6G: motivations, use cases,
Ŕãň¼āãƑƀÙīġġřģąÙ¼ŔąīģŌ˚̝DģD)))Œň¼ģŌ¼ÙŔąīģŌ
requirements, challenges and future directions." 2021
īģŒãň¼āãƑƀ…ÙąãģÙã¼ģߌãÙāģīěīùŸ˛űīě˚ɣ˛
1st IEEE International Online Symposium on Joint
no. 1, pp. 256-263, Sept. 2011, doi: 10.1109/
īġġřģąÙ¼ŔąīģŌΟ…ãģŌąģù˺RΟ…˻˛ɤɢɤɣ˚
TTHZ.2011.2159552.

[20] Kumar Swagat, Anima Majumder, Samrat Dutta, Rekha

[10] P. H. Pathak, X. Feng, P. Hu, and P. Mohapatra.
Raja, Sharath Jotawar, Ashish Kumar, Manish Soni, et
"Visible light communication, networking, and
al. "Design and development of an automated robotic
ŌãģŌąģù˜ŌřƒãŸ˛ŅīŔãģŔą¼ě¼ģßÙā¼ěěãģùãŌ˚̝DģD)))
ŅąÙėΟŌŔīŲŌŸŌŔãġøīň¼ģã̀ÙīġġãňÙãŲ¼ňãāīřŌã˚̝
īġġřģąÙ¼ŔąīģŌ…řƒãŸŌΟŒřŔīňą¼ěŌ˛űīě˚ɣɩ˛ģī˚ɦ˛
¼ň¯ąűŅňãŅňąģŔ¼ň¯ąű˜ɣɩɢɥ˚ɢɤɥɦɢ˺ɤɢɣɩ˻˚
ŅŅ˚ɤɢɦɩ̀ɤɢɩɩ˛;īřƑāŇř¼Ƒãňɤɢɣɧ˛ßīą˜ɣɢ˚ɣɣɢɫ˫
d\…Œ˚ɤɢɣɧ˚ɤɦɩɨɦɩɦ˚ [21] H. Zhan et al. "Evaluating PM2.5 at a construction
ŌąŔãřŌąģùŔãň¼āãƑƀň¼ßą¼Ŕąīģ˚̝DģD)))Œň¼ģŌ¼ÙŔąīģŌ
[11] DŒ‘̀Œ˚̝ŒňřŌŔąģDŒ˚̝ɤɢɣɩ˾dģěąģã˿˚ű¼ąě¼Øěã˜āŔŔŅŌ˜˫˫
īģŒãň¼āãƑƀ…ÙąãģÙã¼ģߌãÙāģīěīùŸ˛űīě˚ɧ˛
ŲŲŲ˚ąŔř˚ąģŔ˫ŅřØ˫Œ̀Œ‘Œ̀Œ‘…Œ̀ɤɢɣɩ
no. 6, pp. 1028-1034, Nov. 2015, doi: 10.1109/
[12] J. Joung, C. K. Ho, K. Adachi, and S. Sun. "A ŒŒA¸˚ɤɢɣɧ˚ɤɦɩɩɧɫɨ˚
ŌřƒãŸīģŅīŲãň̀¼ġŅěąƍãň̀ÙãģŔňąÙŔãÙāģąŇřãŌ
[22] dřŅãģùVąãŔ¼ě˚̝DģŔãùň¼ŔãßŌãģŌąģù¼ģß
øīňŌŅãÙŔňřġ̀¼ģßãģãňùŸ̀ãƉÙąãģŔŲąňãěãŌŌ
communication in 6G: A prototype of high resolution
ÙīġġřģąÙ¼ŔąīģŌ˚̝DģD)))īġġřģąÙ¼ŔąīģŌ…řƒãŸŌ
ŒAƀŌãģŌąģùīģŅīƑ¼ØěãßãűąÙã˚̝)ř^ɤɢɤɣ˚
ΟŒřŔīňą¼ěŌ˛űīě˚ɣɩ˛ģī˚ɣ˛ŅŅ˚ɥɣɧ̀ɥɥɥ˛;ąňŌŔŇř¼Ƒãň
2015, doi: 10.1109/COMST.2014.2350018. [23] T˚…Ŕ¼Ŕģąėīű˛R˚<ňƀŸØ˛˚Aãąģãġ¼ģģ˛¼ģß‘˚˚~øãąƆãň˚
"160-GHz to 1-THz multi-color active imaging
[13] Neil C. Thompson, Kristjan Greenewald, Keeheon
with a lens-coupled SiGe HBT chip-set." In IEEE
Vãã˛¼ģß<¼Øňąãě;˚\¼ģŌī˚̝ŒāãÙīġŅřŔ¼Ŕąīģ¼ě
Œň¼ģŌ¼ÙŔąīģŌīģ\ąÙňīŲ¼űãŒāãīƔ¼ģߌãÙāģąŇřãŌ˛
ěąġąŔŌīøßããŅěã¼ňģąģù˚̝¼ň¯ąű˜ɤɢɢɩ˚ɢɧɧɧɪűɣ˾ÙŌ˚V<˿
vol. 63, no. 2, pp. 520-532, Feb. 2015, doi: 10.1109/
10 Jul 2020.
Œ\ŒŒ˚ɤɢɣɦ˚ɤɥɪɧɩɩɩ˚
[14] C. Zhang, G. He, Y. Chen, P. Zhu and S. Dou.

29 | Huawei Technologies
Technology
[24] ˚Ař¼ģùãŔ¼ě˚̝Aīěīùň¼ŅāąÙ\D\dŌřƐ¼ÙãŌøīňɨ<ŲąňãěãŌŌ
ģãŔŲīňėŌ˜dŅŅīƑřģąŔąãŌ˛Ùā¼ěěãģùãŌ˛¼ģßŔňãģßŌ˚̝DģD)))
ªąňãěãŌŌīġġřģąÙ¼ŔąīģŌ˛űīě˚ɤɩ˛ģī˚ɧ˛ŅŅ˚ɣɣɪ̀ɣɤɧ˛
October 2020, doi: 10.1109/MWC.001.1900534.

[25] Tataria H, Haneda K, Molisch A F, et al. "Standardization

of propagation models for terrestrial cellular systems: A
historical perspective[J]." International Journal of Wireless
Information Networks, 2021, 28(1):20-44.

[26] Z. Yu, Y. Chen, G. Wang, W. Gao, and C. Han. "Wideband

channel measurements and temporal-spatial
¼ģ¼ěŸŌąŌøīňŔãň¼āãƑƀąģßīīňÙīġġřģąÙ¼ŔąīģŌ˚̝ɤɢɤɢ
IEEE International Conference on Communications
Workshops (ICC Workshops), 2020, pp. 1-6, doi: 10.1109/
ICCWorkshops49005.2020.9145314.

˾ɤɩ˿ J. He, Y. Chen, Y. Wang, Z. Yu, and C. Han. "Channel

measurement and path-loss characterization for low-
Ŕãň¼āãƑƀąģßīīňŌÙãģ¼ňąīŌ˚̝ɤɢɤɣD)))DģŔãňģ¼Ŕąīģ¼ě
Conference on Communications Workshops
(ICC Workshops), 2021, pp. 1-6, doi: 10.1109/
DªīňėŌāīŅŌɧɢɥɪɪ˚ɤɢɤɣ˚ɫɦɩɥɩɧɨ˚

[28] 3GPP. "3GPP TR38.811 - Study on New Radio (NR) to

ŌřŅŅīƑģīģ̀ŔãňňãŌŔňą¼ěģãŔŲīňėŌ˚̝©ãňŌąīģɣɧ˚ɦ˚ɢ˛dÙŔ˚ɤɢɤɢ˚

Huawei
HuaweiTechnologies
Technology | 30