A Prospective Look: Key Enabling Technologies, Applications and Open Research Topics in 6G Networks

The fifth generation (5G) mobile networks are envisaged to enable a plethora of breakthrough advancements in wireless technologies, providing support of a diverse set of services over a single platform. While the deployment of 5G systems is scaling up globally, it is time to look ahead for beyond 5G systems. This is mainly driven by the emerging societal trends, calling for fully automated systems and intelligent services supported by extended reality and haptics communications. To accommodate the stringent requirements of their prospective applications, which are data-driven and defined by extremely low-latency, ultra-reliable, fast and seamless wireless connectivity, research initiatives are currently focusing on a progressive roadmap towards the sixth generation (6G) networks, which are expected to bring transformative changes to this premise. In this article, we shed light on some of the major enabling technologies for 6G, which are expected to revolutionize the fundamental architectures of cellular networks and provide multiple homogeneous artificial intelligence-empowered services, including distributed communications, control, computing, sensing, and energy, from its core to its end nodes. In particular, the present paper aims to answer several 6G framework related questions: What are the driving forces for the development of 6G? How will the enabling technologies of 6G differ from those in 5G? What kind of applications and interactions will they support which would not be supported by 5G? We address these questions by presenting a comprehensive study of the 6G vision and outlining seven of its disruptive technologies, i.e., mmWave communications, terahertz communications, optical wireless communications, programmable metasurfaces, drone-based communications, backscatter communications and tactile internet, as well as their potential applications. Then, by leveraging the state-of-the-art literature surveyed for each technology, we discuss the associated requirements, key challenges, and open research problems. These discussions are thereafter used to open up the horizon for future research directions.


I. INTRODUCTION
The phenomenal growth of connected devices and the increasing demand for high data services have been the main driving forces for the evolution of wireless technologies in the past decades. A forecast study reported by the International Telecommunication Union demonstrates that the volume of mobile data will continue to grow at an exponential rate, reaching up to a remarkable figure of about 5 zettabytes per month in 2030 [1]. Meanwhile, due to the emergence of the Internetof-Everything (IoE) paradigm, supporting smart homes, smart cities, and e-health applications seamlessly through connecting billions of people and devices over a single unified communication interface, there is an urgent need to shift the focus from the rate-centric enhanced mobile broadband services to ultra-reliable low-latency communications (URLLC) in order to provide a networked society through massive machine-type communications (MTC) [2], [3]. Besides generating massive data, the upsurge of IoE will naturally give rise to a myriad of new traffic and data service types, leading to diverse communication requirements. This grand vision requires a radical departure from the conventional "one-size-fits-all" network model of fourth generation systems.
The fifth generation (5G) of wireless technology represents a technological breakthrough with respect to the previous communication networks. In addition to reducing latency, enhancing connectivity and reliability, and achieving gigabits per second speeds, 5G is set to deliver a variety of service types, often characterized by conflicting requirements and diverse sets of key performance indicators, simultaneously over one platform [4]. These features make 5G a key enabler for the Internetof-Things (IoTs) application environments, where machineto-people (M2P) communications (e.g., industry automation, smart cities, and intelligent mobility) and machine-to-machine (M2M) communications (e.g., autonomous communications between sensors and actuators) are expected to take place alongside people-to-people communications, (e.g., voice over internet protocol (IP), video conferencing, video streaming, and web browsing). Delivering a plethora of services, with profound differences in terms of quality of service (QoS) requirements, poses major challenges, such as the need to manage a huge volume of a mixture of human-type and machine-type traffic, which is heterogeneous in nature. To cater to these unique challenges, 5G deployment trends to adopt two main network functionalities, namely softwarization and virtualization [5], [6]. By jointly exploiting softwarization and virtualization, cognition and programmability of the end-to-end network chain may be achieved by decoupling the network functions from the hardware platform. This yields enhanced flexibility and reliability, as well as fast network auto-reconfiguration, enabling a larger portfolio of use cases and applications to be supported concurrently.
In parallel with addressing the aforementioned challenges, 5G has introduced potential disruptive technologies to meet stringent requirements in terms of capacity, connectivity, communication resilience, reliability, deployment costs, power consumption, latency, and data rate. These technologies include, but are not limited to, millimeter wave (mmWave) communications, massive multiple-input multiple-output (MIMO), non-orthogonal multiple access (NOMA), and network ultradensification [7], [8].
Despite the strong belief that 5G will support the basic MTC and URLLC related applications, it is arguable whether the capabilities of 5G systems will succeed in keeping the pace with the rapid proliferation of new IoE applications, which are expected to increase by 12% yearly, and which are enabled by massive connectivity and are based on data-centric and automated processes. Meanwhile, following the revolutionary changes in the individual and societal trends, in addition to the noticeable advancement in human-machine interaction technologies, the market demands by 2030 are envisaged to witness the penetration of a new spectrum of IoE services. These services span from extended reality, which comprises augmented reality (AR), virtual reality (VR), and mixed reality services, to flying vehicles, haptics, telemedicine, autonomous systems, and human-machine interfaces. The unprecedented requirements imposed by these services, such as delivering ultra-high reliability, extremely high data rates, and ultralow latency simultaneously over uplink and downlink, will push the performance of 5G systems to its limits within 10 years of its launch, as speculated in [9]. Moreover, the emergence of such new IoE services necessitates integrating the computing, control, and communication functionalities into a single network design.
To deliver future cutting-edge services and accommodate their aforementioned heterogeneous requirements, a new breed of challenges have to be addressed. Examples of these challenges include leveraging sub-terahertz (THz) bands, governing the network performance set by a targeted rate-reliabilitylatency trade-off, provisioning flexibility in the network architecture and functionalities, and designing an intelligent holistic orchestration platform to coordinate all network resource aspects, including communication, control, computing, and sensing, in an efficient, self-sustainable, and scalable manner, which is tailored to the demands of a specific application scenario or use case [10]- [13].
The evolution of 5G has urged the conceptualization of beyond 5G (B5G) wireless systems, including the sixth generation (6G), which be capable of unleashing the full potentials of abundant autonomous services comprising past, as well as emerging trends. More precisely, 6G is envisioned to bring novel disruptive wireless technologies and innovative network architecture into perspective. In particular, 6G is anticipated to: • Offer capacity expansion mechanisms to address the massive scale connectivity aspect and to provide ultrahigh throughput, even in extreme or emergency scenarios where varying device densities, spectrum and infrastructure availability, as well as traffic patterns may exist. • Achieve the targeted quality of immersion and per-user capacity and offer a unified quality of experience required by AR and VR applications, which will hit retail, tourism, education, etc. • Deliver real-time tactile feedback with sub-millisecond (ms) latency to fulfill the needs of haptic applications, such as e-health. • Incorporate artificial intelligence (AI) to support seamless data-centric context-aware communications for controlling environments such as smart structures, autonomous transportation systems, and smart industry. • Meet the extremely high levels of communication reliability (e.g., more than 99.9999%) and the low end-to-end latency to support ultra-high mobility scenarios such as flying vehicles.
While 5G services have begun to roll out across markets, interest in 6G trends has already gained significant momentum both in academia and industry. Several research studies have appeared in the recent literature reporting key technological trends and new research directions that would bring 6G into reality, for example, see [9]- [14]. In [9], the authors presented a speculative study on the main use cases that are expected to be brought by 6G and discussed their associated challenges and the potential enabling technologies. The authors in [10] presented a vision of some potential 6G applications and trends, and discussed the associated service classes and their performance requirements. Additionally, they briefly listed their enabling technologies and pointed out some key open research avenues. In [11], the authors presented an overview of a number of potential 6G revolutionary technologies and the associated network architectural innovations that are envisioned to address the shortcomings of 5G systems. The authors in [12] delivered a roadmap towards enabling AI in 6G. In particular, they discussed key AI methodologies that can play a central role in the design and optimization of 6G networks. In [13], the authors highlighted that in order to support future use cases, current communication infrastructure has to evolve at both physical and architectural levels. They also discussed the need to develop mechanisms to enable a holistic resource management platform and described the resulting challenges in terms of privacy and security. Finally, the authors in [14] presented a vision of 6G and its requirements. With respect to users' perspectives, they also identified innovations that need to be considered towards realizing this vision.
It is noted that the aforementioned reported contributions mainly take a rather use cases-centric approach to the roadmap of 6G era with a focus on the associated services and technological trends. Conversely, in this survey, we approach the 6G vision from the angle of enabling technologies that manifest themselves as the paradigms needed for the realization of 6G. Specifically, we present an in-depth conceptual overview of the main revolutionary technologies in a holistic manner, taking into explicit account the key drivers, performance metrics, and major ongoing research for every single technology. Apart from the technologies discussed in the previous surveys, which are reviewed here in detail, we shed light on additional innovative technologies, such as THz communications, metasurfaces (also known as intelligent reflective surfaces), backscatter communications (BackCom), tactile internet (TI), and aerial networks, which are envisaged to promote the 6G revolution. This survey also delves into the emerging applications of each technology and identifies their associated challenges. This discussion is used to provide a directional guidance towards future research work.
The remainder of the article is organized as follows: In Section II, we present a comprehensive overview of five disruptive 6G technologies. This is followed by outlining some of their potential applications in Section III, whereas Section IV highlights the fundamental challenges associated with each technology discussed in Section II. Finally, the article is concluded in Section V.

II. KEY ENABLING TECHNOLOGIES FOR 6G NETWORKS
Future 6G systems will require the support of novel technologies to enable unprecedented functionalities in the network. These technologies are envisioned to introduce a plethora of new applications associated with remarkably stringent requirements in terms of latency, reliability, energy, efficiency, and capacity, compared to their 5G counterparts. In this section, we provide a concrete conceptual background of major disruptive technologies that will shape the future 6G networks, which includes THz communications, programmable metasurfaces, drone-based communications, BackCom, and TI.

A. Terahertz Communications
One of the key challenges towards realizing 6G networks is the scarcity of spectrum, owing to the unprecedented broadband penetration rate and the emergence of new use cases with rigorous bandwidth requirements. A promising solution to the current spectrum crunch is to explore the THz-band, which is envisioned to bridge the gap between the mmWave band and infrared light-waves (optical communications), by providing a considerably wider bandwidth and enabling the development of new use cases with high data rates requirements. In addition to extending the bandwidth, THz communications offer an amplified gain due to the shorter wavelength experienced at these bands, allowing for the deployment of a large number of antennas.
On the other hand, THz based communications require rethinking of existing solutions and investigate novel approaches that offer a seamless operation over the entire THz band. For example, the design of efficient beamforming and tracking techniques that are able to dynamically and precisely track down the location of THz-enabled devices is of great importance, and an open research problem. Other open issues include hardware architecture design and the integration of massive MIMO and intelligent surfaces. An overview of the opportunities and challenges associated with THz communications is given in [15].
1) State-of-the-Art: Motivated by the important role of THz modulators in enabling THz technology in future wireless systems, the performance of several amplitude and phase modulators was examined for various materials and fabrication processes [16]- [23]. For example, silicon (Si) substrates, coated with effective materials such as gold [16] and manganese iron oxide [17], as well as graphene-based modulators [24], [25] are proven to offer a performance enhancement to the THz modulators by extending the transmission range and pumping power density, in addition to their tunability and high-speed characteristics. Nevertheless, the main drawbacks of these nano-particle-based THz modulators are the high cost and increased complexity. On the other hand, graphenebased modulators suffer from a low modulation depth and high energy consumption.
The implementation of THz communications in outdoor environments is rather challenging, which is particularly due to the inevitable loss caused by molecular absorption and other atmospheric conditions, such as rain [26]. Accordingly, THz transceiver and antenna designs have to be thoroughly investigated prior to the effective design and deployment of these systems. To this end, antenna and transceiver designs have attracted great interests in the research community [27]- [40]. Specifically, the development of Si-germanium signal generators, quantum cascade laser photonic sources, compact graphene antennas and graphene/liquid crystal based phase shifters are some of the reported research work in the field of THz transceiver and antenna designs [41].
Moreover, channel modeling is vital in THz communications to ensure reliability and high spectral efficiency. Existing research is focusing on the characterization of line-of-sight (LOS) and non-LOS (NLOS) components, with emphasis on scattering properties for the NLOS component and free space loss, molecular absorption and harsh weather conditions for the LOS component. In particular, research efforts are mainly centered around characterizing channel coefficients for deterministic and statistical conditions in indoor and outdoor environments. Ray-tracing is a reliable method for modeling LOS and NLOS components and is utilized extensively to characterize the deterministic and stochastic channel coefficients. For instance, in [42]- [48], the authors proposed efficient propagation deterministic models for THz nano-communications while incorporating the LOS and NLOS components for 2D and 3D scenarios. Although deterministic models provide higher accuracy in describing channel coefficients compared to stochastic models, the underlying high computational complexity and the required geometrical information of the propagation environment are critical drawbacks of such models. On the other hand, statistical characterization of the THz channels is rather challenging, especially when taking into account channel mobility, channel state information estimation, and channel correlation. Recent advancements in the design of THz communications systems are summarized in Table II.

Design aspects
Refs.
THz modulator [16]- [23] Antenna design [27]- [34] Transceiver design [35]- [40] Channel modeling [42]- [56] B. Programmable Metasurfaces for Wireless Communications The mmWave and THz communications are envisioned as key enablers for 6G systems. They are expected to satisfy the stringent requirements of various potential 6G use cases by exploiting higher frequency bands. However, owing to the severe attenuation and scattering properties, the detrimental effects on communication efficiency remains the grand challenge in wireless communications. For example, small and large objects in indoor environments, e.g., walls and furniture, typically scatter rays in all directions, leading to severe multipath propagation environments. The Doppler effect is another key challenge, which can limit the realization of ultra-broadband communications, particularly in the mmWave and THz bands [57]. Existing solutions mainly rely on device-side approaches, which consider the wireless environment to be uncontrollable and hence, it remains unaware of the on-going communication processes.
1) Metasurfaces: Metasurfaces have recently emerged as an innovative technology, which is envisioned to revolutionize wireless communications by allowing wireless system designers to fully manipulate the propagation of electromagnetic (EM) waves in a wireless link. The building block of a metasurface is a meta-atom, which is an artificial conductive structure with engineered EM properties that is repeated periodically across a rectangular surface (also called tile). At the macroscopic level, metasurfaces exhibit unique EM properties such as customized permittivity and permeability levels, and negative refraction [58]. As a consequence, metasurfaces enable unprecedented capabilities when interacting with impinging EM waves, which include wave focusing, absorption, imaging, scattering, polarization, to name but a few [58]. It is worth mentioning that metasurfaces leverage these unique abilities without any limitation on their operating frequency.
Recently, there has been a steadily growing interest in both industry and academia on tunable metasurfaces, also called programmable metasurfaces. In this context, the meta-atom design can be dynamically altered through a simple external stimuli, such as a binary switch, empowering metasurfaces with unique adaptivity. More specifically, dynamic meta-atoms are fitted with tunable switching components, such as microelectro-mechanical switches or CMOS transistors, which can alter the structure of the meta-atoms. This allows metasurface tiles to receive commands from an external programming interface, where parameters of the incident and reflected waves, e.g., phase, amplitude, frequency, and polarization, are carefully manipulated in order to enable the EM behavior of interest [57]. Moreover, the discovery of communicating nodes in the surrounding wireless environment can be realized by equipping the metasurface tile with efficient sensing and reporting features.
Tunable mechanisms of metasurfaces facilitate massive connectivity, interference mitigation, and enhanced diversity by introducing an additional degree of freedom. These tunable features are essential in order to realize the flexibility needed for future wireless communications. The authors in [59] presented the first model to describe a programmable wireless indoor environment using programmable metasurfaces. The introduction of programmatically controlled wireless environments has undeniably opened the door for a broad range of functionalities to be achieved even at the mmWave and THz bands. Some metasurface functionalities are summarized as follows: − Beam steering: This function can be achieved by allowing a metasurface to change the direction of the impinging wave towards the desired direction through manipulating either the refraction or reflection index which can override the outgoing directions defined by the Snell's Law [57]. − Beam splitting: In this function, a metasurface tile splits an incident wave into customized orthogonal multiple beams steered towards multiple directions to serve multiple users simultaneously [59]. − Wave absorption: Blocking the access of an unauthorized wireless device can be accomplished by adjusting the properties of the metasurfaces to ensure no or minimal reflection or refraction of the incident wave. This functionality can be utilized to prevent eavesdropping and optimize the network physical-layer security. As demonstrated in [60], for a given incident wave, metasurfaces are able to reduce the wave power by 35 dB. − Wave polarization: This function allows a metasurface to fully control the polarization of impinging waves and manipulate their oscillation orientation [59]. − Phase control: This functionality of metasurfaces allows for the alteration of the carrier phase [59].
2) State-of-the-Art: Programmable metasurfaces have attracted the attention of the research community recently. In particular, a number of research studies examined the potentials of programmable metasurfaces as modulators [61]- [65]. Furthermore, the research works in [66]- [69] investigated the design of smart beamforming in metasurface-based wireless secrecy communication systems. Researchers also explored the use of metasurfaces in wireless power transfer (WPT) [70]- [72]. The role of machine learning in controlling the functionalities of metasurfaces to actively improve the coverage of the highly dynamic indoor environments is analyzed in [73]- [75]. The aforementioned state-of-the-art is summarized in Table  III.

C. Drone-Based Communications and Autonomous Systems
A key driving force behind the vision of 6G is the deployment of connected and autonomous vehicle (CAV) systems and drone-based communications. The research efforts in CAV

Addressed Schemes
Refs.

Modulation [61]-[65]
Beamforming [66]- [69] Wireless power transfer (WPT) [70]- [72] Machine learning [73]- [75] and drone (also known as unmanned aerial vehicles (UAVs))based communication systems, have been steadily growing in both academia and industry, targeting strict requirements, particularly ultra-low latency and unprecedented communication reliability. The advantages, categories, applications, and challenges of drone-based systems are depicted in Fig. 1. In the following, we focus our attention on the current and futuristic application scenarios of UAVs, as well as the state-of-the-art. An in-depth discussion of the underlying challenges will be provided in the next section. 1) Drone-assisted Wireless Communications: Drones constitute the basic building block of aerial networks, whose inherent features, such as mobility and flexibility, enable several imminent and futuristic applications in wireless networks [76]. It is shown that the use of drones can significantly improve the coverage and transmission rates [77]. Furthermore, standardization activities led by 3GPP are currently ongoing to adapt the necessary changes in order to integrate drones in the future wireless networks [78].
Drone-assisted wireless communications can be categorized as follows: Cellular-enabled Drones (CED): CEDs are operated as user equipments (UEs) (i.e., drone-UEs) in order to enable several key applications such as mining, oil and gas, transportation, surveying and monitoring, with a velocity of 160 km/h in urban and rural environments [79]. To ensure connectivity with cellular networks, several essential requirements, which include reliable and low-latency communications between the drone-UEs and ground base stations (BSs), have to be met. In some applications, drone-UEs will require high-speed connectivity from ground BSs or from drone-BSs (i.e., drones operating as BSs). It should be pointed out that nowadays cellular networks are designed for ground users with unique mobility and traffic characteristics that are considerably distinct from those experienced with drone-UEs. Therefore, the integration of drone-UEs into cellular networks in a single wireless network presents a set of new key challenges and design considerations that must be addressed for the efficient realization and successful deployment of CEDs [80].
Wireless Infrastructure Drones (WIDs): WIDs are intended to extend the network capabilities by enhancing network coverage or capacity. WIDs can be further classified based on their functionalities into: • Drone-BSs: Drone-BSs are aerial nodes with some BS features and functionalities, that are envisioned to provide capacity and coverage enhancements for 6G networks. They are cost-effective solutions that render wireless connectivity to hard-to-reach areas, as well as geographical areas with limited cellular infrastructure. Drone-BSs are also attractive solutions for delivering reliable, broadband and wide-scale temporary wireless connectivity in special events or harsh scenarios, such as sport events and natural disasters. Furthermore, high altitude drone-BSs are expected to provide a long-term and cost effective connectivity for rural areas. The integration of drone-BSs with other physical layer techniques such as mmWave and massive MIMO, cognitive radios, etc., is a promising solution to provide data-hungry services and is expected to create a new set of challenges for the next generation of flying BSs [80]. Optimal positioning of the drone-BSs is one of the critical challenges that needs to be addressed in dense deployment scenarios [81]. • Aerial Relays: Relaying has been extensively investigated in the context of terrestrial communications to enhance network reliability, throughput, and coverage. However, such relays are subject to limited mobility and often are constrained by wired backhauling [82]. On the contrary, drones acting as wireless relays, are versatile and offer high mobility. This feature makes them a promising candidate for providing enhanced wireless connectivity beyond LOS. Moreover, aerial relays can play a significant role in extending the battery life of drones [83]. • Aerial Backhaul for Cellular Networks: Wireless backhauling has been shown to provide a cost-effective solution compared to wired backhauling. However, it is subject to interference, blockage and path loss, which can significantly degrade the performance and reduce the data rate [84]. In this respect, drone-based networks are foreseen to play a fundamental role in realizing robust and high-speed backhaul connectivity for cellular networks [85]. Such networks are expected to provide flexible drone-based backhaul communications that will enhance the network capacity, reliability, as well as the operation cost [80]. 2) Air-Ground-Space Integrated Networks: There is a strong belief that existing terrestrial, aerial and satellite networks will not be able to cope with the volume of generated data, which will continue to grow at an exponential rate together with the rapid proliferation of new IoE applications. On the other hand, the integration of these networks is seen as the next evolution of wireless infrastructure, that is envisioned to cater to diverse use cases with different QoS requirements, particularly in realistic scenarios such as urban, rural, and lightly dense areas. Yet, despite their indispensable benefits, the envisioned integrated architecture will introduce unprecedented challenges that include, but are not limited to, heterogeneity, security, resource management, self-organization, energy consumption, and backhauling [86].
3) Drone-based Multi-access Edge Computing: Multiaccess edge computing (MEC), enables cloud computing capabilities at the edge of cellular networks, and has recently emerged as one of the potential technologies for 5G networks. In particular, MEC enables mobile devices with limited resources to offload their computation tasks to the edge of the network.
In drone-enabled networks, mobile devices can offload their computationally demanding tasks to drones with MEC capabilities, typically at the edge of the network, thus reducing the network congestion and allowing for the rapid deployment of new applications. Additionally, the drone-based network can provide an effective mobility management without the necessity of handover, as well as uninterrupted MEC services for high mobility users, due to their large-scale coverage and LOS connection [87], [88].
Within the same context, the limited on-board processing capabilities of drones, which are mainly due to their storage and battery constraints, bring about several concerns towards the efficient execution of complex tasks [89]. In particular, heavy computation demanding applications, such as real-time image processing, may not be supported by the anticipated vision of drones. Recent research efforts proposed efficient techniques to tackle these limitations by leveraging cloud computing to offload the computation-intensive tasks from the drones to remote cloud servers [90]. The role of these servers can be summarized as follows [91]: − Storage: Storage services can be offered by the cloud to store drones data streams that include environment and mission-related parameters, captured images and sensed data. − Computation: Intensive computations are executed in the cloud in order to minimize the processing time and energy consumption at the drone. Moreover, the large amount of stored data from the drones can be exploited to perform data analytic tasks in order to enhance the performance of drones-enabled networks, in terms of trajectory adaptation, altitude optimization, and energy consumption customization, in an intelligent manner. 4) State-of-the-Art: The research efforts have focused thus far on basic and more advanced design aspects of drones, including CEDs and WIDs. Specifically, extensive research efforts are directed towards proposing encryption and authentication mechanisms to ensure a robust secure communication infrastructure that also maintains privacy [92]- [102].
The energy efficiency and battery properties are key practical design aspects, in which the battery life, charging mechanism and energy consumption must be optimized to enable seamless and uninterrupted wireless communications. In this context, the optimization of energy consumption and charging time of drones have received significant attention [103]- [110].
Various design issues, that allow for the realization of the full potential of drone networks, are tackled in the literature, such as network architecture [86], [111]- [113], image processing [114]- [116], interference management [117]- [122] and storage [123], [124]. Recent advances in aerial networks are summarized in Table IV. The exponential growth of connected devices, constituting the emerging IoE, is a major driving force towards the development of energy-efficient solutions to sustain wireless communication among connected nodes [125]. Nonetheless, despite the notable advancements, the short battery life of the deployed battery-operated devices still constitutes a major design challenge, which requires a paradigm shift towards the development of the next generation green communication architecture. Accordingly, ambient BackCom have emerged as a new communication paradigm for low power communications in 5G networks. This approach is based on the concept that a transmitter sends data to its receiver by backscattering ambient signals, e.g., TV or Wi-Fi signals. Compared to conventional systems, backscatter transceivers consume significantly less power (orders of magnitude), rendering it a strong candidate for low power networks and IoE applications [126]. Owing to its promising features, several new and disruptive technologies can be integrated with BackCom.
1) Radio-frequency (RF)-Powered BackCom Networks: RF energy harvesting (RF-EH) has been recently proposed as a promising solution to provide perpetual energy replenishment for such networks. RF-EH is realized by allowing wireless devices, equipped with dedicated EH circuits, to harvest energy from either ambient RF signals or dedicated RF sources. It can be divided into two main categories, namely wireless-powered communications [127] and simultaneous wireless information and power transfer, which have been shown to provide noticeable gains in terms of power and spectral efficiencies by enabling simultaneous information process. Despite the remarkable advantages, RF-EH techniques still suffer from particular limitations, especially in the context of low power wireless networks. Specifically, wireless-powered devices are not able to communicate perpetually, as they require dedicated time for energy harvesting. Additionally, these devices depend on active RF signals for communication; as a consequence, they suffer from relatively high power consumption, which can pose major issues, particularly in large-scale low power wireless networks [126]. Motivated by this concern, a new trend is to integrate BackCom systems with various RF-EH techniques in a single network. This promising paradigm is envisioned to address some of these challenges and catalyze the deployment of new technologies and services [128].
2) RF-Powered Cognitive Radio Networks and Ambient BackCom: The integration of RF-EH techniques with cognitive radio (CR) networks has led to the development of a new communication paradigm, called RF-powered CR networks [129]. In such networks, a CR transmitter harvests RF energy when a primary (licensed) user (PU) is active, which is subsequently utilized for data transmissions between secondary (unlicensed) users (SUs) [129], [130]. Evidently, the performance of these networks greatly depends on the availability of PU signals. In this context, BackCom is envisioned as a potential solution to address this challenge by allowing SUs to harvest energy from PU signals in addition to transmitting data by backscattering the PU signals. Therefore, it is evident that although BackCom and energy harvesting have not played a major role yet in 5G, they are envisaged to be part of 6G with full potentials.
3) Visible Light BackCom: Visible light communications (VLC) is a new paradigm that is foreseen to provide ubiquitous connectivity while addressing some of the limitations and challenges of RF communications. It is based on intensity modulation and direct detection, where the intensity of light-emitting diodes (LEDs) is modulated to carry information, and then demodulated/detected directly using a photodiode. There are several key advantages of VLC that include inherent communication security, high degree of spatial reuse, and its immunity to RF interference, which makes it safe to be used in critical places with high electromagnetic interference, e.g., hospitals and industrial plants. The principle of visible light BackCom (VLBC) systems is similar to its RF counterpart, in which VLBC leverages ambient light to harvest energy and then modulates VLC signals to transmit its data to backscatter receivers [131], [132]. 4) Quantum BackCom: Quantum backscatter communications is another promising technology which is anticipated to contribute towards the development of 6G and the next generation IoT, particularly in terms of performance and security [133]. In this new paradigm, a transmitter produces entangled signal-idler (S-I) photon pairs. The S-photon is transmitted and backscattered from a backscatter transmitter, while the I-photon is kept at the receiver. This quantum setting provides a significant gain in the error exponent for the communication link and facilitates secure communication by exploiting quantum cryptography.
5) State-of-the-Art: BackCom networks are subject to critical security threats, such as eavesdropping and jamming. This stems from the simplicity and low-complexity of BackCom transceivers. As a result, existing security solutions, including encryption and digital signatures, may not be applicable due to the power and complexity constraints of BackCom devices. This has motivated the research community to investigate new security mechanisms that can guarantee fully secure and private wireless communications [134]- [141]. Self-interference is another major limitation in BackCom systems. The sources of self-interference include: (i) signals from ambient RF sources, and (ii) multipath propagation. In recent works, several selfinterference cancellation techniques are proposed in the literature [142]- [149].
BackCom networks are not optimized and/or designed for large-scale low power networks comprising a massive number of IoT devices, e.g., sensors in environmental monitoring, sensors in smart roads to collect data about the pavement conditions and traffic, etc. Furthermore, such systems are different from human-centric communications with diverse and unique traffic characteristics as well as QoS requirements, which requires the development of efficient physical layer and media access control schemes to prevent access congestion. Within this context, multiple access techniques in BackCom systems are regarded to be instrumental in improving the efficiency of backscatter networks. In particular, conventional orthogonal multiple access and non-orthogonal schemes (e.g., NOMA [150] and rate splitting multiple access), are recognized as promising candidates for enabling massive connectivity, while maintaining high energy and spectral efficiency [151]- [159].

E. Tactile Internet
Tactile Internet (TI) is seen as the next frontier of IoE, focusing on M2P and M2M interactions. With the recent advances in tactile/haptic devices, it is predicted that TI will catalyze the deployment of a plethora of new applications ranging from health care to education and smart manufacturing. Therefore, it is expected to reshape our daily lives and realize the full potential of the next industrial revolution, also known as Industry 4.0.
To fully realize IT, the communication infrastructure (CI) has to meet strict design guidelines, as it is currently unable to address the stringent requirements of the use cases envisioned for TI. In particular, the CI has to support extremely low endto-end latency with high-reliability [176]. Furthermore, it must ensure data security without jeopardizing the latency requirements imposed by the computationally demanding encryption techniques.
To address these requirements and catalyze the deployment of new use cases with unique requirements, the development of unique and disruptive B5G wireless communication technologies is of paramount importance. To this end, we envision the development of: 1) communication technologies in the THz band; 2) novel network architectures; and 3) AI-enabled networks.

A. Terahertz Communications
The wide bandwidth in the THz band is envisioned to drive the deployment of a large array of new use cases, as outlined next.
• Wireless Data Centers: Today's data centers suffer from high complexity, power consumption, maintenance cost, and wasted spaces occupied by large cables. Therefore, there have been attempts to address these issues in order to enable fast and reliable access to cloud-based services. According to [213], the power consumption of all data centers will reach 73 billion KWh by 2020. Therefore, THz communications could be a promising candidate for the next generation of data centers, satisfying the peak data rate of 10-20 Gbps required by 5G, and even higher [214]. Although still in infancy, there have been recent research investigations on channel modeling for indoor environments, which have paved the way for utilizing THz communications in indoor wireless data centers [215], [216]. Initial results showed that the cabling cost can be reduced without compromising the bandwidth. • Secure Drone Communications: Drone communications in the THz band is one of the envisioned applications of THz communications that are expected to achieve higher capacity gains and support increased mobility [217]. Moreover, the deployment of large antenna arrays for coverage extension enables extremely narrow beams, which inherently limits the probability of eavesdropping, yielding secure communications [218]. • Health Monitoring: THz communications is a promising candidate in the field of health care. Specifically, several nano-sensors can be utilized to monitor different ions in the human blood, such as glucose and sodium, in addition to cholesterol levels, infections and cancer bio-markers. The collected data by the sensors are forwarded to a micro interface, e.g., a cellular phone or a medical device, using THz communications [213]. It is noted that the THz radiations are considered safe for the human-being bodies compared to the gamma rays [219]. • Wireless local area networks (WLANs)/Wireless personal area networks (WPANs): THz band communications are envisioned to enable bandwidth-intensive applications such as high definition holographic video conferencing and ultra-high speed data transfer. This stems from the fact that a seamless interconnection may be facilitated between ultra-high wired networks (e.g., fiber optical links) and wireless devices (e.g., laptops or tablets) in WLANs or between personal wireless devices in WPANs [218].
Potential applications of THz communications are outlined in Fig. 2.

B. Metasurfaces for Wireless Communications
• Metasurfaces in WPT Applications: WPT is foreseen as a game-changing technology, in which future networks are envisioned to provide perpetual energy replenishment, particularly for low power devices/sensors. A major concern, however, is the ability of devices to harvest enough energy in wireless channels. The unique properties of metasurfaces, that include their abilities to steer and concentrate electromagnetic waves, enable efficient power transfer and energy harvesting. The work in [220] discussed the potentials of integrating smart tables with metasurfaces in order to enable multiple wireless devices to be powered/charged simultaneously. The integration of WPT in metasurfaces for biological applications was studied in [71], where a metasurface-based wearable device was placed over the human skin surface to improve the efficiency of an implanted WPT system. • Metasurface-based Textiles for Wireless Body Sensor Networks (WBSNs): Very recently, metasurface-based textiles were developed for energy-efficient and secure WBSN applications [221]. In this approach, regular clothing is fitted with conductive metasurface textiles, where wireless signals can glide around the surface of the body on the clothes to interconnect wireless wearable devices with each other forming a WBSN. In this application scenario, wearable devices are located in close proximity to the body. This results in a significant reduction in the power dissipated by the wireless devices, leading to an improvement in the battery life and data rates of these devices. In fact, this innovative WBSN is foreseen to boost the received signal compared to conventional technologies. Furthermore, metasurface-based textiles may enable personal sensor networks, which are highly efficient, immune to interference, and inherently secure [221]. Looking ahead, they are envisioned to have future applications in high-tech athletic wear, health monitoring, and humanmachine interfaces.

C. Drone-Based Communications and Autonomous Systems
• Search and Rescue Missions: Search and rescue missions are some of the critical driving applications of drone networks. This is primarily due to the flexibility of drones compared to manned vehicles, which take a longer time to deploy [222]. • Mailing and Delivery: Package delivery is one of the attractive civil applications of drones, adopted by major courier companies around the world in order to accomplish fast, cost-effective and reliable delivery. This is motivated by the fact that most of the packages' weights are below the maximum tolerable load of a single drone [89]. For example, Amazon reported that 83% of their packages weights fall below the 2.5 kg [223], while FedEx average package weight is less than 5 kg [89]. performing intensive studies relating to marine organisms and ecosystem [224].

D. BackCom and Energy Harvesting
• Smart Homes: Low power battery-less backscatter sensors equipped with energy harvesting devices can be efficiently embedded in homes to perform a wide range of tasks, such as gas leak detection, smoke and carbon oxide detection, and movement monitoring. Another driving application of BackCom is smart dustbins, in which backscatter devices keep track of the garbage level and report it to garbage collecting trucks. • Smart Cities: Backscatter enabled sensors can be flexibly placed in street lamps, parking lots, buildings, and bridges to realize the envisioned energy-efficient low-cost smart cities. BackCom can be utilized in smart cities to enhance air quality by monitoring the pollution and noise level in the air. Additionally, it can be used to manage traffic in closed parking areas and ease the process of finding an available parking place by indicating the available slots. • Biomedical Applications: Wearable and implantable human medical devices, in addition to plants and animals monitoring, are some of the key drivers of BackCom technology. For example, Smart Google contact lenses, which are equipped with miniaturized BackCom devices, are designed to continuously measure the glucose levels in the tears for diabetes patients and backscatter the reported results to a wireless controller. Other serious diseases, such as epilepsy and Parkinsons, are envisioned to be diagnosed and treated by the assistance of Back-Com technology. In particular, it is envisaged that brainimplantable BackCom neural devices will play the role of the brain-computer interface needed for studying and diagnosing diseases of interest.
Potential applications of BackCom systems are presented in Fig. 3.

E. Tactile Internet
• Industry: Automation in industry realizes the control of machinery and processes through a large network of sensors and actuators in order to improve productivity and reduce labor costs. Industrial automation is steadily growing in the context of TI, enabling the full control of rapidly moving devices with high sensitivity while meeting the end-to-end latency requirements. However, the ever-growing need for control processes with different latency, reliability, data rate, and security demands is envisioned to catalyze the development of new wireless solutions tailored to these requirements. • Virtual Reality (VR): VR enables users to physically interact with each other by applying various motor skills over a VR simulation platform. In this context, TI is anticipated to provide the low latency required to facilitate shared virtual environments. High-fidelity interaction requires haptic feedback to allow users to touch objects in a VR environment and enable users to feel one another's actions on the same touched object. This requires a stable and seamless user communication coordination, which is not supported by today's VR systems. Hence, TI is foreseen as a key enabler for haptic communications with ultra-low delay communication and reliability requirements. • Augmented Reality (AR): AR applications are fast growing, owing to the availability of AR glasses and powerful smart devices equipped with small sensors and cameras. However, the present AR systems are restricted to deliver pre-processed content due to the limited computational capabilities of the small wireless devices and the inherent delays in the communication network. TI, on the other hand, is perceived to enable the augmentation of dynamic and real-time information to the contents. • Healthcare: Potential applications of TI in healthcare include tele-surgery, tele-rehabilitation, and tele-diagnosis. Different from healthcare services provided by current communication networks, which are location-dependent, medical expertise provided by the TI will not be bounded by time and/or a physical location. For example, a physician can diagnose patients at their locations by remotely controlling a robot while receiving haptic feedback as well as audio-visual information. Tele-surgery is another example, which has the potential to revolutionize healthcare delivery in the next decade. • Education: Improved learning experiences over distances can be achieved via TI by allowing teachers and learners to exchange haptic information. Identical multi-modal human-machine interfaces are required to enable auditory, visual, and haptic interactions, which can be realized by enabling ultra-low latency communication systems. For example, TI may allow a remote music instructor to apply instant actions over the haptic overlay to correct the hand moves of a student learning a musical instrument.

IV. CHALLENGES AND FUTURE DIRECTIONS
In this section, we discuss the open research issues associated with the previously presented potential 6G technologies and highlight their research challenges.

A. Terahertz Communications
• Transceiver Architecture: Transceiver architectures in the THz band is one of the critical aspects to be considered due to the unique characteristics of the propagation environments of THz links. In order to realize the full potentials of the THz-band, there is a growing interest in the development of novel transceiver architectures that can operate across the entire THz-band. The developed architectures are expected to combat the severe path loss, thus, enabling high sensitivity and high power gains. Moreover, the co-existence of different frequency bands, such as THz, mmWave, and microwave cells, requires thorough investigations over different layers. • THz Modulator: The characteristics of THz modulators, including amplitude and phase modulators, play a central role in quantifying the efficiency of THz communication systems. These characteristics include, but not limited to, modulation speed and depth in amplitude modulators and phase shift amount in phase modulators. Current modulators designs, with the adopted architecture and utilized materials, limit the modulator ability to achieve ultra-high speed and consequently to realize efficient THz wireless systems. This stems from the fact that the existing modulators do not allow EM radiation manipulation in the THz band, which is required in order to facilitate high-speed control of the modulator characteristics. Therefore, this calls for research intervention to develop intelligent and tunable ultra-high speed modulators, with approximately 1 picosecond response time, to enable efficient and reliable THz wireless communications [213]. • Channel Modeling for THz Communications: Existing low frequency channel models can not accurately capture the behavior of high frequency THz links, which experience severe attenuation due to molecular absorption and antenna aperture, in addition to the free space loss.
Note that the multi-path channel of THz communications compromises LOS and NLOS components. On the contrary, LOS attenuation, represented by path loss, is measured by the addition of the spreading and molecular absorption losses, which are encountered due to wave expansion and molecular absorption, respectively. The severity of molecular absorption is determined based on the density of molecules experienced along with the transmission link, distance, weather conditions (e.g., heavy rain), and frequency window in the THz band. Accordingly, LOS channel component in the THz band is described as severely frequency selective. Therefore, it is essential to develop an accurate model to represent the LOS component in the THz wireless system, which is necessary to identify the performance limits of THz communications and propose enhancement schemes for such technology [225].
On the other hand, due to the unavailability of the LOS components in some scenarios, the THz link might be limited to the NLOS component, which can be classified into specular reflected, diffusely scattered and diffracted EM waves. Therefore, for precise channel characterization, it is required to accurately trace the reflection, scattering and diffraction coefficients of the incident beam in the THz system [226], which depend on the incident angle and surface material and geometry. Hence, the development of realistic and accurate channel models for THz links is still an open research problem, which requires thorough investigation to enable the implementation of an efficient THz wireless system.

B. Metasurfaces for Wireless Communications
In spite of the promising prospects of metasurfaces in 6G, several design aspects should be further investigated in order to realize the full potential of this promising technology.
• Dynamic Structure Design: The ability to manipulate the configurations of meta-atoms constitutes a key design challenge for the efficient operation of reconfigurable metasurfaces, whose deployment is needed to support a wide range of functionalities in highly dynamic wireless environments. Although there exist some research studies that have successfully demonstrated that multiple functionalities can be achieved by multiple metasurfaces, only a few have presented the capability of a metasurface to perform different functionalities simultaneously [227]. In this case, each unit cell of the metasurface has to be controlled independently, raising the need to develop effective distributed meta-atom control mechanisms and to examine the performance of the variety of functions supported by each metasurface. Additionally, since metasurfaces are envisioned to be deployed in application scenarios involving the operation over a wide frequency range (varying from 1 to 60 GHz), designing efficient metasurface structures that are capable of dynamically switching the operation frequency poses an essential research goal [58]. • Efficient Programmable Interface: Apart from the need to develop metasurface structures capable of realizing different functions in real-time, there is a compelling need to investigate advanced multi-functional metasurfaces that can switch from one EM behavior to another in a fast manner to cater for the increasingly diverse user demands, especially in high mobility scenarios where the system convergence rate may not be within the coherence time of the surrounding wireless environment. As a result, research efforts should be directed towards developing control software that incorporates low-complexity and fast configuration optimizers to facilitate the optimization and adaptation of metasurfaces functionalities to the surrounding environment. Also, advanced signal processing and machine learning algorithms may be developed to leverage the sensing capabilities of metasurfaces for enabling intelligent system performance optimization, which can converge within the coherence time of the environment and can be aligned with the network requirements of 6G systems, such as massive connectivity, ultra-low latency, and high reliability [58]. • High-order Modulation: The design of high-order modulation and novel waveform designs for metasurface-based wireless communication systems constitute promising solutions for enabling high data rate transmissions. This is of paramount importance since current metasurfacebased transmitters are limited to single-carrier low-order modulation schemes, such as binary/quadrature phaseshift-keying [62], [63]. • WPT in Metasurfaces: It is recalled that the last years witnessed remarkable advancements in battery design. Nonetheless, the short battery life of wireless devices still constitutes a major design challenge and requires a paradigm shift towards the development of the next generation green communication architectures. WPT was proposed recently as a promising solution to provide perpetual energy replenishment for such networks. It is realized by allowing wireless devices, equipped with dedicated energy harvesting circuits, to harvest energy from either ambient RF signals or dedicated RF sources. Given that metasurfaces have the ability to steer, absorb and collimate EM waves, particular research efforts should be dedicated to exploit the unique functionalities of metasurfaces to wirelessly charge the wireless devices from long distances.

C. Drone-Based Communications and Autonomous Systems
• Network Architecture and Analysis: Network planning, performance evaluation and resource allocation are some of the challenges that drone-BSs encounters in aerial networks. While terrestrial mobile networks are designed to meet the requirements of ground users, they are not optimized to support aerial networks. Specifically, terrestrial BS antennas are not designed to support the ultralow latency requirements of high elevation angle users in aerial networks. Therefore, there is a need to develop a novel and efficient system architecture that can efficiently integrate terrestrial BSs with drone-based UEs. • Energy and Storage Efficiency: Energy constraint is a limiting factor in mobile-enabled drones, particularly since solar energy and the limited size of built-in batteries are the only sources of power. This is a crucial issue, especially in power-hungry monitoring missions, where continuous monitoring and transmission are inevitable. Various energy-aware mechanisms have been reported in the literature to address the energy efficiency problem in drones. For example, an approach is to utilize multiple cooperative drones to allow a single drone to temporarily leave the network for energy replenishment [82]. Storage constraint is another major concern, e.g., in monitoring missions, where drones must store a large amount of data. This motivates the investigation of novel forwarding and compression schemes to efficiently handle this huge amount of data. • Collision Avoidance: Buildings and large obstacles represent a major hazard to drones, so they must be addressed thoroughly in order to avoid collisions to objects in the surrounding environment. A way to address this problem is to restrict the drone flying zones to limited areas. However, this will increase the interference between multiple drones and lead to higher collision probability [228]. Therefore, there is a need for efficient collision avoidance schemes to enable drones dynamically adjust their trajectories to minimize collision probability. • Channel Modeling: Efficient implementation of cooperative aerial networks requires accurate characterization of communication links to ensure reliable and safe operation of air-to-air and air-to-ground links. In flying ad-hoc network architectures, drone communications require the development of robust theoretical framework to model air-to-air and air-to-ground links. While there have been reported works on link characteristics of aerial networks in different frequency bands [229]- [231], there are still not enough results to characterize the channel models, particularly for cooperative (relaying) scenarios [232]. Although the communication link characteristics of drone-based systems are unique, some terrestrial channel models, such as two-ray and Rician models, were shown to be a good fit for drone environments; however, more experimental and real-time tests are required in order to verify the accuracy of such models and properly select their parameters. More importantly, further research efforts must be dedicated to verify the validity of these models in different frequency bands, such as 433 MHz, 1575.42 MHz, and 2.4 GHz bands [232].

D. Backscatter Communications and Energy Harvesting
• Security and Jamming: BackCom systems typically suffer from potential security and jamming attacks, owing to their simple modulation and coding schemes. The key issue is that the limited resources in backscatter systems are not able to support the implementation of conventional security solutions that include encryption and digital signatures [142]. This calls for the development of simple, yet highly efficient security solutions to realize secure BackCom systems. • Interference to Licensed Systems: Data transmission in ambient BackCom is based on reflecting ambient signals received from licensed sources. Therefore, interference imposed on licensed users is inevitable, which calls for the need to develop communication protocols that guarantee no or minimal interference. Recent research efforts have focused on interference modeling and development of compensation schemes [126]. • Full-Duplex Ambient Backscatter: Full-duplex BackCom systems are proposed to enable simultaneous communication between multiple ambient backscatter nodes. In such cases, the same antenna is used by a backscatter receiver to transmit and receive signals. As a result, a significant amount of self-interference exists between different components of the BackCom transceiver. This calls for the development of self-interference mitigation schemes and constitutes an open research issue towards addressing this challenge [143], [233].

E. Tactile Internet
Although TI is considered as a new paradigm envisioned to generate a plethora of new applications, several open research challenges exist and need to be fully addressed for the successful realization of this enabling technology.
• Haptic Devices: Haptic devices, such as sensors and actuators, enable users to feel, touch, and manipulate objects in real or virtual environments. Although haptic devices have already been commercialized, they still fall short in terms of degrees of freedom as well as the cost effectiveness. Additionally, in order to realize the envisioned applications, haptic devices have to offer kinesthetic and tactile control simultaneously [234]. • Data Compression: Bandwidth-limited networks represent a major challenge for haptic communications, which requires a paradigm shift towards the development of innovative solutions that would enhance system reliability and user experience. In this regard, several haptic data compression techniques have been thoroughly investigated in the literature, to realize the full potential of TI. However, further investigations towards haptic codec design for TI are required. This might include the proposal of a new set of kinesthetic and tactile codec solutions that will lead to highly efficient data compression techniques. • Integration of Multi-Modal Sensory: One of the key challenging aspects in enabling haptic feedback is multimodal sensory, where visual, haptic, and auditory feedback are integrated simultaneously. However, these different modalities vary in terms of their latency, sampling, and transmission rate. Subsequently, novel multiplexing schemes have to be studied in order to temporally integrate multiple modalities with different priorities. • Ultra-Reliability: Since TI is expected to disrupt major attributes in the society, ultra-reliable network connectivity is necessary to minimize the packet losses and reduce the outage to 10 −7 [234]. A highly lossy environment in haptic communications leads to erroneous sensations and directly interrupts the user's activity. There are several factors that impact the reliability of TI applications. This includes uncontrollable interference, lack of resources, equipment failure, and reduced signal strength. This will require the investigation of efficient reliability enhancement mechanisms to achieve ultra-high reliability in realtime operations [235]. • Ultra-Low Latency: As stated earlier, TI requires sub-ms end-to-end latency. Therefore, it is essential to understand the latency budget between sensors and actuators in order to investigate the impact of each contributing factor in the chain. In general, the end-to-end latency is dominated by air-interface, backhaul, and core latencies. To cater to the critical latency requirements, innovative latency optimization mechanisms are necessary in addition to effective protocol stack and hardware designs.

V. CONCLUSION
Although the glory of 5G networks is at its peak, initial implementation and testing phase of 5G networks along with the emergence of a plethora of new applications, such as bio-interface applications, are revealing new challenges and limitations of the upcoming wireless networks, including but not limited to, ultra-high reliability, extremely high data rates and ultra-low latency. Accordingly, this spots the lights on the fundamental question: Will the forthcoming 5G wireless networks be able to accommodate the newly emerged applications with the concurrent stringent requirements necessary for realizing fully autonomous and intelligent systems? To answer this question, in this paper, we sketched out the roadmap into the future hypothetical vision of B5G networks. Particularly, we focused on exploring expected new technologies for 6G networks, such as THz communications, metasurfaces, aerial networks, BackCom, and TI, along with their potential applications and inherent challenges. The technical challenges associated with these technologies call for a deeper investigation, which will potentially accelerate the development of innovative solutions as well as standardization efforts for 6G.