From a Human-Centric Perspective: What Might 6G Be?

As the standardization of fifth generation (5G) communications has been completed, and the 5G network will be commercially launched in 2020, the research visioning and planning of sixth generation (6G) communications are being initiated. 6G communications are expected to be the next focus in wireless communication and networking and aim to provide remarkable communication services to meet the future demands in the 2030s. We believe that the human-centric mobile communications will still be the most important application of 6G and 6G network should be human-centric. Following this rationale, high security, secrecy, and privacy are the all-important features of 6G, which shall be paid special attention from the wireless research community. To promote the research and the human-centric design ideology for 6G communications, we imagine in this article a comprehensive and systematic framework of 6G with five slices, key features, and enabling technologies. Furthermore, we briefly discuss the issues of 6G beyond communication technologies. Overall, the article aims to provide an envisioned picture of 6G and serve as a research guideline in the post-5G era.


S INCE the initial development of the Advanced Mobile
Phone System (AMPS) by Bell Labs, which was later called the first generation (1G) network, there have been three large-scale and radical updates of wireless communication networks in the past four decades, resulting in the second, third, and fourth generation (2G, 3G, and 4G) networks [1]. The fifth generation (5G) network launching is ongoing and expected to be commercialized by 2020. As the standardization of 5G has gradually been solidified, an increasing number of researchers place their visions on the future sixth generation (6G) communication network and initialize a series of advanced research planning activities [2]- [7].
What might 6G be? Is 6G nothing more than a hype? Will 6G be a technological and commercial success? To answer these questions, we provide our thoughts in this article. Different from the published perceptions, we believe that the conventional mobile communications will still be the most important application of 6G around 2030, although other application scenarios are becoming ubiquitous and increasingly significant. Consequently, 6G network should be human-centric, instead of machine-centric, application-centric, or data-centric. Following this rationale, high security, secrecy, and privacy are the all-important features of 6G. Also, user's satisfaction of service (SoE) would be adopted as a pivotal metric in 6G communication networks, and a number of new features are required accordingly. and tablets, mobile communications became the mainstream and contributed to a considerable amount of data throughput in 4G networks [2]. The 4G era since 2009 till now is undoubtedly the 'Golden Decade' of mobile communications. The information and communications technologies (ICTs) accompanying 4G have greatly reshaped our world societally, economically, culturally, and technologically [10]- [14].

B. What Has 5G Been?
In 2014, an enlightening paper was published, which discusses what 5G will be and points out that the key technologies to achieve 5G are network densification, millimeter wave, and massive MIMO architecture [15]. A magnificent scene of the Internet of Everything (IoE) is commonly described and indulged in the literature of 5G. Five years passed since then, the conception of 5G has gradually been solidified in recent days, and main technological companies and operators have launched their construction plans for 5G networks so as to realize the large-scale commercial deployment by 2020. First year of the 5G era as the all-important theme has been discussed in the Sixth Annual Brooklyn 5G Summit 2019. Therefore, it is now an appropriate moment to examine what 5G has been.
In the first deployment stage of 5G networks, most operators and device manufacturers adopt the 3GPP 5G New Radio (NR) standard for dense urban areas [16]. The corresponding 5G network operates on the 2-6 GHz spectra. Both millimeter wave and massive MIMO technologies are widely used in 5G networks, while the network densification construction is delayed for certain reasons. Network slicing is more or less involved in 5G mission-critical solutions. Internet Protocol television (IPTV) and high-definition (HD) video streaming, service over high-speed mobility, basic virtual reality (VR) and augmented reality (AR) services can be well supported. Indoor services and data services in dense metropolitan areas will continue to be the main focus in the 5G era. For different application scenarios, a complete 5G communication network provides three service options: enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive machine-type communications (mMTC) [17].
On the other hand, there are also a variety of state-ofthe-art communication and networking technologies that have not been incorporated in 5G standards yet. The main reasons can be summarized from two sides. From the supply side, some technologies still require experimental verification and in-depth tests in practical environments. Meanwhile, the high cost and unsatisfactory backward compatibility also prevent them from being used. From the demand side, the services and devices supported by some advanced communication and networking technologies have not been widely on demand.
Although 5G has adopted a gradual evolution strategy that is able to provide much more and better services than 4G, it is admitted that there is no ground-breaking technology in 5G that really bring a surprise to us. Inheriting the fundamental performance enhancement mechanisms since 3G [18], more spectral and hardware resources are invested in 5G to achieve performance gains, instead of a technological breakthrough.
To be informative, we illustrate the perceptual scenarios for a single user from the 1G era to the 5G era and 6G era (hypothetical) in Fig. 1. In 1G and 2G, users perceived by voice and text. In 3G and 4G, picture and video help to construct a colourful perception in mind, respectively. In 5G, live ultra-high-definition three-dimensional data is employed to produce a comprehensive perception. It is highly expected that we will have ubiquitous virtual existence and involvement in the coming 6G era.

C. Current Research Progress Towards 6G
To realize technological breakthroughs in practical communication networks and achieve a new perceptual dimension, the research community of wireless communications has gradually turned attention to and anchor the hope on 6G. The 6G vision and requirements are suggested in [2], in which special attention is paid to the battery lifetime of mobile device and service classes in 6G, rather than data rate and latency. In [3], it is pointed out that the communication system research in the post-5G era must incorporate with circuit and device manufacturing capabilities so as to form a closed feedback loop of research activities. A number of new communication scenarios in future networks around 2030 are predicted in [4], which encompass holographic calls, flying networks, teleoperated driving, and the tactile Internet, etc. Further, it is foreseen that the same level of reliability as wired communications will be offered to future wireless communications. [5] and [6] summarize the future driving applications and trends as well as enabling technologies in 6G networks. In particular, network decentralization based on blockchain technology is believed to be a key to simplify network management and provide satisfactory performance in 6G. The concept of human-centric service is also proposed and viewed as the emphasis in 6G.
The key performance indicators (KPIs) of 6G are defined, and a speculative comparison between 5G and 6G is provided in [7].
For practical implementations, multiple access, air interface and data center for 6G communications are envisioned and discussed in [19], [20] and [21], respectively. Networking patterns of 6G networks are outlined in [22]- [24], in which cell-less architecture, decentralized resource allocation, and three-dimensional super-connectivity are highly expected to exist in 6G networks. MTCs and vertical-specific wireless network solutions for 6G are studied in [25], which believes that 6G would facilitate the first wall-breaking standard to completely replace existing industry-specific communication standards and provide a unified solution enabling seamless connectivity for all needs in vertical industries.
Among all technological works pertaining to 6G, terahertz (THz) communications, artificial intelligence (AI), and reconfigurable intelligent surfaces are the most eye-catching protagonists. They are viewed as paradigm-shifting and revolutionary technologies in wireless communications. A comprehensive study of THz communications for 6G is reported in [26], which includes a detailed technological overview, transmitterreceiver designs, and various practical demonstrations. AI empowered 6G is believed to be able to provide a series of new features, e.g., self-aggregation, context-awareness, selfconfiguration, and opportunistic setup [27]. Additionally, AI empowered 6G would unlock the full potential of radio signals and enable the transformation from cognitive radio (CR) to intelligent radio (IR) [28]. Machine learning is in particular crucial for realizing AI empowered 6G from the algorithmic perspective, which has been detailed in [29]. Besides the algorithms, reconfigurable intelligent surfaces are supposed to be used to construct the hardware foundation of AI in wireless communications [30]. Reconfigurable intelligent surfaces are also envisaged as the massive MIMO 2.0 in 6G and analyzed in [31]- [33]. These attractive materials can also incorporate with index modulation (IM) to yield an increase in spectral efficiency in 6G networks [34].
Apart from above-released works, a number of 6G projects have already been started around the world, which aim to attain the initiative, define 6G, and reshape the framework as well as the business model of wireless communications. The first project refers to the 6Genesis Flagship Program (6GFP), a recently formed Finish consortium, which is followed by Terabit Bidirectional Multi-user Optical Wireless System (TOWS) for 6G LiFi started at the beginning of 2019.
In March 2019, the first 6G Wireless Summit was held in Levi, Finland and formally triggered the starting gun of 6G research race in academia. Besides the summit, a number of small-scale workshops and seminars were also held worldwide to discuss the possibility of 6G, e.g., Huawei 6G Workshop, Wi-UAV Workshop of Globecom 2018, and Carleton 6G Workshop.
Except for academia, 6G and future networks also attract standardizing bodies, industrial organizations, and governments. IEEE launched IEEE Future Network with the tagline 'Enabling 5G and Beyond' in August 2018. ITU-T Study Group 13 also established the ITU-T Focus Group Technologies for Network 2030 intending to understand the service requirements for future networks round 2030. Project Loon was triggered by Google and is now running independently, which plans to provide reliable Internet connection to the unconnected five billion population. A research group based on Terranova is now working toward the reliable 6G connection with 400 Gbit/s transmission capability in the THz spectrum.
LG Electronics also announced the foundation of 6G Research Centre at Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea. Samsung kicked off its 6G research in June 2019. SK Telecom has decided to collaborate with Nokia and Ericsson in 6G research in the mid of 2019. In late 2018, China's Ministry of Industry and Information Technology declared the ambition of leading the wireless communication market around 2030 by expanding the research investment in 6G. Federal Communications Commission of the U.S. opened 95 GHz to 3 THz spectra for the use of 6G research, which marks the participation of the U.S., the world's biggest economic entity, in the 6G research race. In addition, an EU/Japan project under the ICT-09-2017 H2020 called 'Networking Research beyond 5G' also investigates the possibility of using THz spectrum from 100 GHz to 450 GHz.
II. POTENTIAL APPLICATION SCENARIOS AND CHALLENGES 6G communications are expected to provide improved services in terms of coverage, data rate and allow users to connect each other everywhere. It is expected to adopt unconventional communication networks to access several types of data and transmit them through conventional improved radio frequency (RF) networks, allowing new communication experience with virtual existence and involvement anywhere. To explicitly define the probable features of 6G communications, we foresee the potential application scenarios and challenges for 6G communications in this section. It should be noted that as a speculative study of 6G ('What Might 6G Be'), we intend to cover a large range of heated topics discussed in recently published works and conference releases, but with our own thoughts and comments to appraise these 6G candidate technologies.

A. Enhanced Conventional Mobile Communications
As we declared at the beginning, 6G communications should be human-centric, which implies that the conventional mobile communications will still hold the position of protagonist in 6G, in which classic cellular phone is the major tool of mobile communications. The challenges regarding conventional mobile communications comes from five aspects: 1) how to enhance security and protect privacy; 2) how to expand network coverage in a rapid and cost-efficient way, especially in distant and isolated areas; 3) how to reduce the cost of mobile communications; 4) how to extend the battery life of the mobile device; 5) how to provide a higher data rate with a lower end-to-end latency.

B. Accurate Indoor Positioning
With the help of the Global Positioning System (GPS), outdoor positioning becomes full-fledged and can be regarded as accurate in most application scenarios now. However, indoor positioning is still far from maturity, because of the complex indoor electromagnetic propagation environment [35]. Accurate and reliable indoor positioning services will radically change the living habits of mobile users and open up new niches for economic prosperity. On the other hand, there is a growing consensus that accurate indoor positioning might not be viable by sole utilizing RF communications [36]. Such a crucial and impactful application is highly expected to be realized in the era of 6G with more advanced non-RF communication technologies.

C. Communications of Wearable Devices, Integrated Headsets, and Implantable Sensors
In addition to the classic mobile phone and tablet, it is foreseen that there will be an increasing number of new communication devices in the 2030s. These new communication devices can be wearable devices, integrated headsets, and implantable sensors [37]. Different from the portable phone and tablet, these emerging devices impose diverse environmental and system requirements on communication networks. For example, there must be strict constraints on transmit power and frequency band used in these devices for health reasons. The device weight will become more sensitive when designing wearable devices and integrated headsets. Reliable power supply and security for implementable sensors are of high importance. In addition, there should be major dissimilitudes in mathematical modeling between these emerging communication devices and classic mobile phones and tablets.

D. High-Quality Communication Services on Board
Despite the effort and endeavour of researchers in the 4G and 5G eras, it is undeniable that communication services on board are still unsatisfied in most cases nowadays. The communication services provided on board are challenging by the high mobility, Doppler shift, frequent hand-over, lack of coverage, and so on [38]. Satellite communications enable communication services on board with acceptable service quality, but are too costly, especially in aircraft cabins [39]. To provide high-quality communication services on board, not only new communication technologies must be employed in 6G communications, but also novel networking architectures shall be in use.

E. Worldwide Connectivity and Three-Dimensional Integrated Networking
In the last decade, researchers drew attention to the communication services in dense metropolitan areas, especially for indoor communication scenarios. However, it should not be omitted that there is a large population around the world having no access to basic data services, especially in sparse, developing, and rural areas [40]. The advent of the astonishing 6G era should not only benefit the majority in dense areas, but be shadowed to less dense areas. Making wireless networks not only vertical but also horizontal would benefit a much larger population. In this regard, worldwide connectivity is expected to be realized in 6G communications by a low-cost implementation scheme in order to guarantee the communication fairness of minority in sparse areas. Providing this service is greatly dependent on novel networking technologies.
To achieve the goal of worldwide connectivity, threedimensional integrated networking would be utilized, which encompasses terrestrial, airborne, and satellite communications [41]. Apart from satellite communications, most existing communication and networking architectures only consider two-dimensional scenarios, in which the heights of communication nodes are negligible [42]. This modeling assumption is appropriate and efficient for 5G application scenarios.
However, it is envisioned that communications of flying nodes for achieving worldwide connectivity become ubiquitous in the 2030s and shall be taken into consideration when planning 6G networks. Such a three-dimensional integrated network could bring considerable performance gains and unprecedented services to users [5].
Apart from the communications on ground and over sky, extending the communication network to underwater environment is a crucial and even necessary element of worldwide connectivity, especially that more than 70% of the earth's surface is covered by water, and several marine applications needs live monitoring [43]. Underwater optical wireless communication (UOWC) can play a vital role in establishing reliable high data rate links with the help of acoustic communications [44]. Underwater communication nodes such as autonomous vehicles, sensors and divers can be connected by underwater BSs using UOWC. Moreover, underwater communication networks are connected to terrestrial networks via water surface networks and aerial networks. Securing a sustainable energy source by wind, sun and water flow is an essential requirement for both underwater and water surface networks.

F. Communications Supporting Vertical Industries
To deeply serve the physical world, 6G communications are also expected to continuously support the applications in vertical industries, including building and factory automation, manufacturing, e-health, transportation, agriculture, surveillance, and smart grid. These applications are essential to the Industry 4.0 and believed to be true paradigm shifts [45]. They pose special service requirements in addition to conventional mobile telephony and broadband data communications. In particular, these vertical industries normally necessitate high standards with respect to connection reliability, transmission latency, and security [46]. To integrate these vertical industries in 6G communications, mMTC in 5G needs to be upgraded. More MTC application scenarios and types of machine nodes are required to be considered.
In general, these vertically industrial applications can be classified into robotic communications and vehicular communications. Robotic communications are related to the communications of kinesthetic robotics and manufacturing robotics. Because any error, delay and malicious action in robotic communication applications could result in severe instability, robotic communications are reliability-critical, delay-critical and security-critical. Moreover, a huge number of heterogeneous data streams are generated in robotic communication networks, which yield a challenge for the current centralized networking architecture [47].
For vehicular communications, two emerging technological trends in the vehicular industry are reshaping the physical world, which are corresponding to the self-driving and remotedriving technologies [48]. Due to the development of both technologies in recent years, it is believed that they will be technically mature and widely applied before 2030. To enable both driving technologies in practice, massive vehicleto-everything (V2X) communications must be studied and incorporated in 6G, which provide the basis for high-reliability and low-latency as well as secure exchange of massive driving and ambient data.

G. Holographic Communications
6G is expected to be a conversion point from the traditional video conferencing to a virtual in-person meeting. To this end, a realistic projection of real-time movement needs to be transferred in negligible time, which resorts to holographic communications [49]. In fact, transferring three-dimensional image along the voice is not sufficient to convey the in-person presence. There is a need to have a three-dimensional video with stereo audio that can be reconfigured easily to capture several physical presences in the same area. In other words, one can interact with the received holographic data and modify the received video as needed. All this information needs to be captured and transmitted over reliable communication networks that should have an extremely large bandwidth.

H. Tactile Communications
After using holographic communication to transfer a virtual vision of close-to-real sights of people, events, environments, and etc. It is beneficial to remotely exchange the physical interaction through the tactile Internet in real time [50]. Specifically, the expected services include teleoperation, cooperative automated driving, and interpersonal communication, where it should be possible to apply haptic control through communication networks. Efficient cross-layer communication system design has to be conducted to meet these stringent requirements of these applications. For example, new physical layer (PHY) schemes need to be developed, such as to improve the design of signaling systems, waveform multiplexing, and etc. As for the delay, all delay sources should be treated carefully, including buffering, queuing, scheduling, handover and the ones induced from protocols. Existing wireless communication systems cannot fulfill these requirements, and there is a necessity to over-the-air fiber communication systems [51].

I. Human Bond Communications
6G is expected to widely support the human-centric communication concept, where the human can access and/or share physical features. Human bond communication concept is proposed to allow accessing the human five senses [52]. Recently, the concept is expanded with the help of 'communication through breath' scheme to allow reading the human bio-profile using the exhaled breath and even interact with the human body by inhalation using volatile organic compounds [53]. As a result, it is possible to diagnose diseases, detect emotions, collect biological features and interact with the human body in a remote way. Developing communication systems that can replicate the human senses and human biological features requires interdisciplinary research. It is expected to have hybrid communication technologies that are able to sense different physical quantities and then share it with the intended receiver in a secure manner.

J. Summary
Summarized from the above application scenarios, we boldly anticipate five slices supported by 6G communications: These five slices are pictorially illustrated in Fig. 2. We explain each of them in the following paragraphs.
1) Enhanced Mobile Broadband Plus (eMBB-Plus): The eMBB-Plus in 6G is the successor of the eMBB in 5G, serving the conventional mobile communications with much higher requirements and standards. It should also be more capable of optimizing the cellular networks in terms of interference, hand-over, as well as big data transmission and processing. Additional functionality will also be provided with an affordable expense to subscribers, e.g., accurate indoor positioning and globally compatible connection among diverse mobile operating networks. Most importantly, special attention of security, secrecy, and privacy shall be paid to the eMBB-Plus communication services.
2) Big Communications (BigCom): Different from 5G that emphasizes extremely good communication services in dense areas but to some extent neglects the service in remote areas, the BigCom in 6G cares about the service fairness between dense and remote areas. To be feasible, the BigCom does not intend to provide equally good services in both areas but keep a better resource balance. At least, the BigCom guarantees that the network coverage has to be large enough so as to provide acceptable data service wherever the communication subscribers are living or moving to. The Gini index and the Lorenz curve could be involved to evaluate the service fairness provided by the BigCom and should be treated as crucial indicators of SoE in 6G [54].
3) Secure Ultra-Reliable Low-Latency Communications (SURLLC): The SURLLC in 6G is a joint upgrade of the URLLC and the mMTC in 5G, but with higher requirements of reliability (higher than 99.9999999%, i.e., 'Seven Sigma' from the viewpoints of quality control and process improvement) and latency (less than 0.1 ms) [25], as well as an additional demand on security. The SURLLC mainly serves the industrial and military communications, e.g., a variety of robots, highprecision machine tools, and conveyor systems in the 6G era. In addition, vehicular communications in 6G could also greatly benefit from the SURLLC. 4) Three-Dimensional Integrated Communications (3D-InteCom): The 3D-InteCom in 6G stresses that the network analysis, planning and optimization shall be raised from two dimensions to three dimensions, by which the heights of communications nodes must be taken into consideration. Satellite UAV, and underwater communications can be the examples of this three-dimensional scenario and benefit from threedimensional analysis, planning and optimization. Accordingly, the analytical framework constructed for two-dimensional wireless communications stemmed from stochastic geometry and graph theory needs to be updated in the era of 6G [55]. Considering the node height also enables the implementation of elevation beamforming with full-dimensional MIMO architectures, which provides another direction for network optimization [56].

5) Unconventional Data Communications (UCDC):
The UCDC is probably the most open-ended slice in 6G communications. We intend to propose this slice to cover those novel communication prototypes and paradigms that cannot be classified into another four slices. Currently, the definition and embodiment of the UCDC is still awaiting further exploration, but it should at least cover holographic, tactile, and human bond communications.

III. KEY FEATURES AND ENABLING COMMUNICATION TECHNOLOGIES OF 6G
Based on the application scenarios and challenges as well as the five supported slices in 6G discussed in the last section, we are now able to define the key features of 6G in this section. To enable the key features of 6G, multiple state-ofthe-art communication technologies must be jointly applied, which are also summarized in this section.
To begin with, a qualitative comparison between 5G and 6G communications is summarized in Fig. 3. In this figure, we first suppose that the spectral efficiency in 5G has already been close to the boundary by the advances in massive MIMO, network densification, and millimeter-wave transmission as well as a set of legacy multiplexing techniques inhering from 4G. As bounded by the Shannon limit, the spectral efficiency in 6G would hardly be improved on a large scale. In contrast, security, secrecy, and privacy in 6G communications should be significantly enhanced by new technologies. In 5G networks, traditional encryption algorithms based on the Rivest-Shamir-Adleman (RSA) public-key cryptosystems are still in use to provide transmission security and secrecy. The RSA cryptosystems have become insecure under the pressure of Dig Data and AI technologies, let alone privacy protection mechanisms that are far from being full-fledged in the 5G era. Incremental improvements would happen for energy efficiency, intelligence, affordability, and customization. The energy efficiency gain would be accomplished by the maturity of energy harvesting technology and green communications. Intelligence in 6G can be classified into operational, environmental, and service levels, which will benefit from the thrust in AI developments. The improvements on affordability and customization rely on novel networking architectures, promotion and operational strategies on the market. To be more specific, we give detailed comparisons from 1G to 6G communications in Table I.

A. High Security, Secrecy, and Privacy
Researchers placed great emphasis on network throughput, reliability, latency, and the number of served users in 4G and 5G communications. It has also been widely recognized that the two most efficient ways to improve these metrics are to densify the network and use a higher frequency to transmit signals [18]. However, the security, secrecy, and privacy issues of wireless communications have been, to some extent, overlooked in the past decades. To protect data security, the classic encryption based on RSA algorithms is being challenged by increasingly powerful computers [57]. PHY security technologies and quantum key distribution via visible light communications (VLC) would be the solutions to the data security challenge in 6G [58]- [60]. More advanced quantum computing and quantum communication technologies might also be deployed to provide intensive protection against various cyber attacks [61]. Meanwhile, communication/data service providers have legally collected an enormous amount of user information, and private data leakage incidents happened occasionally. This becomes an unstable factor in the humancentric 6G network and could lead to a disastrous consequence without proper countermeasures. To solve this problem, it is envisioned that complete anonymization, decentralization, and untraceability can be realized in 6G networks by blockchain technology [62].

B. High Affordability and Full Customization
Again, from a human-centric perspective, technological success should not directly or indirectly increase the financial burden or deprive users' options. Therefore, high affordability and full customization should be two important technological indicators of 6G communications. The former is always ignored in most existing works. One might find their proposed solutions/schemes having a much higher transmission rate and/or reliability, but the cost rendered by such improvements will completely restrict their implementations in real life. The academic research activities for 6G should always try to get rid of such speciousness and endeavor to provide high affordability.
Full customization allows users to choose the service modes and adjust individual preference. For example, a user would like to have a low-rate but reliable data service; another user tolerates unreliable data service in order to get a lower communication expense in return; some might only care about the energy consumption of their devices; some even intend to get rid of smart functionality due to the concerns of data security and privacy. All users will be granted the right to choose what they like in 6G, which should not be deprived by intelligent technologies and unnecessary system configurations. Accordingly, the performance analysis of 6G communication systems should also integrate multiple performance metrics into a whole, instead of treating them independently. SoE would be explicitly defined and adopted as a pivotal metric for performance evaluation in the 6G era. That is, unlike 1G-5G, for which we added more elements in the quality of service (QoS) vector, we should map all required performance metrics as a whole to a unique SoE performance metric for each individual user in 6G.

C. Low Energy Consumption and Long Battery Life
The daily charging requirements of smartphones and tablets become annoying in 4G/LTE networks and will continue to the foreseeable 5G era. To release the daily charging constraint for most communication devices and facilitate communication services, low energy consumption and long battery life are two research emphases in 6G communications. To lower energy consumption, the computing tasks of a user device can be off-loaded to smart base stations (BSs) with reliable power supply or pervasive smart radio space [63]. Cooperative relay communications and network densification would also help to reduce the transmit power of mobile device by reducing the per-hop signal propagation distance [64], [65]. To achieve a long battery life, various energy harvesting methodologies would be applied in 6G, which not only harvest energy from ambient radios, but also the energy from micro-vibrations and sunlight [66]. Long-distance wireless power charging would also be a promising approach to extend battery life, but indepth investigations are indispensable to avoid health related issues [67].

D. High Intelligence
The high intelligence in 6G will benefit network operations, wireless propagation environments, and communication ser- vices, which refer to operational intelligence, environmental intelligence, and service intelligence, respectively. 1) Operational intelligence: Conventional network operation involves a great number of multi-objective performance optimization problems subject to a series of complex constraints. Resources, including communication devices, frequency bands, transmit power, and so on are required to be arranged in a proper way so as to achieve a satisfactory level of network operation. Moreover, these multi-objective performance optimization problems are usually NP-hard, and optimal solutions are hard to be obtained on a real-time basis. With the development of machine learning techniques, especially deep learning, a BS equipped with graphics processing units (GPUs) or the control center of core network could carry out relevant learning algorithms to allocate resources efficiently to achieve performance close to the optimum [68].
2) Environmental intelligence: Meanwhile, by the advances on smart radio space and smart materials, distributed and pervasive intelligence of the holistic communication environment, including wireless channels, would become possible [69]. This could provide self-organizing and self-healing properties for the 6G network and enable reliable device-to-device (D2D) communications in a fully intelligent way. Latest works have defined and justified the conception of reconfigurable intelligent surfaces, which are designed to sense the wireless environment and apply customized transformations to the radio waves in an adaptive manner [30], [70]. This conception solidifies the hardware foundation of environmental intelligence.
3) Service intelligence: Furthermore, as a human-centric network, the high intelligence of 6G network also directly reflects in a plethora of communication services, e.g., indoor/outdoor positioning, multi-device management, information search, e-health, surveillance, cyber security [36], [71]. Service intelligence enables these services to be provided in a satisfactory and personalized way. For example, the accuracy of indoor positioning can be greatly improved by deep learning techniques [72], and personalized healthcare is realized by intelligent IoT and multi-model data collecting infrastructure [73]. The service intelligence mainly benefit from high-performance core networks implemented in 6G [74], [75].

E. Extremely Larger Bandwidth than 5G
The THz band defined from 0.1 THz to 10 THz was known as a gap band between the microwave and optical spectra [76]. Nowadays, electronic, photonic and hybrid electronic-photonic based methods are developed [77]. Thus, hybrid THz/free space optical (FSO) systems are highly expected to be realized in 6G using the hybrid electronic-photonic transceivers, where an optical laser can be used to generate THz signal or send an optical signal. The hybrid link offers plenty of bandwidth for signal transmission and has the immunity to adverse weather conditions [78]. THz transmission can play a vital role in the uplink, because a line-of-sight link is not required. Thus, THz uplink solution offers a reliable communication link for VLC networks compared with the infrared solution that needs tracking and positioning system. Hybrid VLC/THz system introduces robust communication solutions against ambient light that reduces the signal-to-noise ratio of the VLC system.

F. Trade-offs Between Key Features and Potential Solutions
It should be noted that as an engineering system, it is not possible to satisfy all wished features without investing more resources, because there exist a number of trade-offs between these features. For 6G communications, we must figure out a way to invest adequate resources in guaranteeing some critical features and raise up all features with an equilibrium.
To this end, we discuss a set of crucial trade-offs in 6G communications regarding these key features and the potential solutions in the following paragraphs.
1) Privacy vs. intelligence: As a human-centric network, the trade-off between privacy and intelligence would be the most important one in 6G communications. On the one hand, AI algorithms need to get access to personal data and process them, so as to optimize network operations, adapt network coefficients, and provide high-quality services. On the other hand, privacy would be sacrificed for the sake of high intelligence. A potential solution is to introduce an intermediate agent between the end-user data and AI algorithms. Such an intermediate agent should be third-party and unmanned if possible, and operate on a decentralized basis. All private and sensitive data will be anonymized by this third-party agent and become untraceable in any way.
2) Affordability vs. intelligence: High intelligence introduces a high degree of system complexity, which could raise up the costs to network operators and device manufacturers. All these raised costs will finally be transferred to less affordable products to end users. To resolve this trade-off, technological breakthroughs in intelligent systems are necessary, but more importantly, a new commercial strategy would be helpful. Once security, secrecy, and privacy are guaranteed, end users have the right to exchange the accessibility of their anonymized data for a lower data price. A similar feature of the smart grid, by which electricity users can also sell selfgenerated electricity back to electricity companies would be borrowed to 6G communication networks.
3) Customization vs. intelligence: High intelligence provided by AI algorithms and smart devices weakens the free will of human beings. In other words, the user preference might not be always aligned with the optimized option produced by AI algorithms. The contradictory situation becomes severer when multiple users are taken into account. This conflict can be formulated as the trade-off between customization and intelligence in 6G communications. In our opinion, the priority shall always be given to the customization, and prohibitive clauses are on demand for AI algorithms and smart devices. These prohibitive clauses should be stipulated in the most fundamental and underlying protocols of 6G communications. In this way, intelligent services can only be provided within the permissible boundary.
4) Security vs. spectral efficiency: Conventionally, to ensure a secure transmission, more spectral resource shall be in use for preventive measures, and the net load for transmitting information is lowered accordingly, given a limited radio spectrum. We have to recognize that this trade-off between security and spectral efficiency is difficult to resolve, but we can mitigate it in three possible ways. First, researchers might try to design a more efficient encryption algorithm. However, this direction would be rather difficult due to the maturity of data encryption. Second, researchers might resort to PHY security technologies for providing security protection without a great loss of spectral efficiency. Third, AI algorithms can also help to detect network anomaly and would be utilized in 6G networks to provide an early warning mechanism for security enhancement. 5) Spectral efficiency vs. energy efficiency: The trade-off between spectral efficiency and energy efficiency is a frequent topic in the field of wireless communications. The discussion pertaining to this well-known trade-off ran through all wireless generations and will, of course, be one of the focuses in 6G communications. However, different from 1G-5G, a new pathbreaking technology would be introduced to greatly alleviate this trade-off, which is energy harvesting. By energy harvesting, user devices are capable of harvesting radio, vibratory, and solar energy from the ambient environment and the constraint on energy consumption can thereby be released. The environmental intelligence realized by ubiquitous intelligent surfaces would also help to mitigate the spectrum-energy trade-off by adapting radio propagation environments.

IV. BEYOND THE COMMUNICATION TECHNOLOGIES
Communication technologies are crucial, but not all. To promote a new technological paradigm and make it socioeconomically profitable, we must always keep the issues beyond technology in mind. In this section, we briefly discuss several crucial issues vis-à-vis 6G beyond the communication technologies per se.

A. Dependency on Basic Sciences
It is admitted that the advancement of wireless communications is highly restricted by basic sciences, especially mathematics and physics. As detailed in [18], we are squeezing the last lemon juice of Shannon's treatise published in 1948 and almost reach the hard wall set by information theory. What is worse, incapable mathematical tools prevent us from exploring the exact performance of communication systems and make us lost in the asymptopia. As a result, a large number of impractical assumptions are made in order to make analysis mathematically tractable, which cannot provide much insight and guideline for 6G communications. The breakthrough in mathematics would often result in a new research boom in wireless communications, and one example is the stochastic geometry and graph theory applied for wireless network modeling [55]. To summarize, researchers shall pay sufficient attention to basic sciences and interdisciplinary fields in order to realize 6G networks.

B. Dependency on Upstream Industries
In the wireless communication research community, it is widely agreed that the most efficient ways to enhance wireless communication systems are to expand the usage to highfrequency spectrum and to reduce the coverage of a single cell [18]. The former tendency is witnessed by the evolution from cellular radio spectrum to millimeter-wave spectrum, THz spectrum, and visible light spectrum. The later tendency refers to the network densification. On the other hand, both tendencies must match up the developments in upstream industries, e.g., electronics manufacturing. First of all, in theoretical research, one can assume an arbitrarily high frequency for use, but in reality, the communication devices constituted realistic electronic components must be able to meet these requirements. In some cases, the resultant data rate has even exceeded the allowable constraint on the electronic circuit by the current manufacturing level, or the signal on higher frequency bands cannot be proceeded by currently commercial chips at all. Neglecting the dependency on upstream industries will turn the 6G research to be nothing but a theoretical carnival.

C. Demand-Oriented Research Roadmap
It has been noticed that there exists a visible mismatch of PHY research activities in industry and academia [11]. As suggested in [3] and [18], a closer connection between industrial and academic researches should be constructed so as to form a positively closed feedback loop for adjusting research roadmap. More directly, such a positively closed feedback loop can be extended to the market and the end beneficiaries of 6G. In this way, a demand-oriented research roadmap can be well designed and adapted in a much more effective and efficient manner. To achieve this goal, it is required to introduce the ideas of value engineering to plan academic research activities. In this way, 6G research roadmap should not be defined by the technological embodiments, but by the function and cost as a whole from a value engineering view. In other words, the research activities in 6G should not simply aim at adding more functions without considering the cost and demand from the end beneficiaries' perspective, but target the value of the implemented service. Specifically, end beneficiaries shall be granted the right to have their voices to reshape the research roadmap in the 6G era. To well satisfy the demands of multiple stakeholders and bridge between academic and real-world problems, the barrier among various disciplines should be removed. More economic and sociological methodologies, e.g, empirical analysis and PES-TEL analysis (PESTEL: an acronym that stands for Political, Economic, Social, Technological, Environmental and Legal factors), could be introduced for appraising and tailoring 6G research roadmap.

D. Business Model and Commercialization of 6G
Previous research activities primarily focus on the technology itself, but rarely on the business model and commercialization. Omitting the marketing aspects would lead to failure (3G could be to some extent an example of such failure [2]). Network densification is a promising solution to satisfy the data transmission burst, but who should pay for it? Building new BSs is costly after all, because of land use right granting and construction operations. Moreover, as 6G communications would bring ground-breaking communication technologies relying on novel network architectures, how to ensure the backward compatibility of 6G with 4G/LTE, Wi-Fi and 5G is still questionable and worth investigating. The overall cost for updating the existing infrastructures for 6G communications needs to be evaluated first, and then the business model and commercialization of 6G should be studied for its commercial triumph. One should always remember that for most ordinary users and government policy makers, paying several times higher to get a dispensable performance gain in terms of transmission rate or latency will not be accepted, let alone appreciated.

E. Potential Health and Psychological Issues for Users
The 'base station myth' is a frequent topic in public media and could even trigger severe protests [79], which reflects the health concerns of users about radiation safety. As reported, electromotive force (EMF) limits have been reached in some cities. With a densified network with a smaller coverage per BS and the use of higher frequency band, there are reasons to believe that such concerns will be aggravated in the era of 6G. As 6G communications are expected to be human-centric, special attention must be paid to the potential health issues brought to users. In this context, EMF-aware transmission would be a novel concept to be introduced in 6G to mitigate the health concerns [80]. Bandwidth expansion from the millimeter-wave regime to the THz regime also causes uncertain biological impacts on humans and animals. Careful studies are required to examine the safety of THz radiation [26].
Apart from health issues, the psychological barrier would also be a factor hindering the large-scale implementation of 6G from a human-centric perspective. As visioned in some proposals even for 5G networks, massive sensors are deployed all over the space, and they are smart to detect, understand, communicate, and respond (fortunately, such a sci-fi scene has been greatly exaggerated). Then, the question will be: will people really enjoy and be comfortable to live in such a smart space? Will people be delighted to be recorded and watched by such a technocratic 'big brother'? Without a careful study on these psychological issues before implementing in practice, 6G could cause catastrophic consequences and even deconstruct existing trust in ICTs [81]. 6G is expected to be not only technologically trustworthy [82], but also societally trustworthy.

F. Social Factors Hindering the Worldwide Connectivity
As pointed out in the background paper of the World Economic Forum at Davos Annual Meeting 2017 [40], apart from technological and economic factors, social factors could also prevent worldwide connectivity in 6G. That is, the people living in developing areas are not motivated to be connected, because of the lack of content relevance, language barrier, and computer literacy. This is mainly a demand-side issue and shall be given full consideration when deploying 6G networks for worldwide connectivity. Incentive schemes and campaigns sponsored by local governments and private companies would be beneficial to persuade the unconnected in distant areas to be connected and promote the concept of worldwide connectivity in the 6G era. The promotions of e-payment and e-taxi in China are good examples that connect most people who never use smart phone before.

V. CONCLUSION
In this article, we provided our basic thoughts on 6G communications from a human-centric perspective. Overall, 6G is not only an extension of 5G with a much higher rate and lower latency, but also should be human-centric. In this regard, we believe that high security, secrecy, and privacy are the allimportant features of 6G, which shall be paid special attention from the wireless research community. We also presented the potential application scenarios of 6G and summarized five slices that should be supported in 6G. Following these slices, key features and enabling technologies for 6G communications were introduced. Beyond the communication technologies, we briefly discussed other crucial issues regarding 6G communications, including the dependency on basic sciences and upstream industries, research roadmap planning, business model and commercialization, as well as health, psychological, and social impacts. The article aims to provide an envisioned picture of 6G and serve as a research guideline in the post-5G era.

ACKNOWLEDGMENT
We thank the editor and the anonymous reviewers for the constructive comments, which have helped us improve the quality of this article. Fig. 1, Fig. 2, and Fig. 3 were created by Ivan Gromicho, Scientific Illustrator at King Abdullah University of Science and Technology (KAUST).