Addressing spectrum efficiency through hybrid-duplex UAV communications: Challenges and opportunities

Abstract

Unmanned aerial vehicles (UAVs) are set to play a prominent role in next-generation communications networks. However, the shortage of available spectrum to support UAV communications is a crucial challenge towards realizing multi-UAV networks. To this end, a hybrid-duplex (HBD) UAV communication system, i.e., HBD-UCS, is a potential solution towards improving spectrum efficiency in UAV communications. In an HBD-UCS, UAVs with half-duplex (HD) transceivers simultaneously communicate on the same spectrum with full-duplex (FD) ground stations (GSs). However, HBD-UCSs are also limited by self-interference (SI) at the FD-enabled GSs and inter-UAV interference at the downlink UAVs. As such, accurate modeling of HBD-UCSs that realistically accounts for SI and inter-UAV interference is necessary before any realization of practical HBD-UCSs is possible. In this regard, this survey paper first discusses the state-of-the-art and open research problems in UAV channel modeling. Next, the literature on the various types of SI mitigation, and the modeling of FD transceiver impairments are discussed. An example signal model for the FD-enabled GS that considers the various FD transceiver impairments is also proposed, before critical technical and regulatory challenges are presented. At the HD UAVs, interference management strategies that can be potentially adopted are discussed, together with the associated state-of-the-art. Furthermore, realistic performance metrics, and open research problems, for the interference management strategies are also discussed. For multi-UAV scenarios in HBD-UCSs, potential non-orthogonal multiple access (NOMA) techniques for multi-UAV deployment are discussed, with an example power-domain NOMA signal model proposed for a NOMA-aided multi-UAV HBD-UCS. Relevant open research problems in power-domain NOMA are then presented for successive interference cancellation (SIC) detection, user-pairing, and DL-based techniques for a power-domain NOMA-aided multi-UAV HBD-UCS.

Introduction

In future generations of communication networks, unmanned aerial vehicles (UAVs) are expected to play an increasingly important role in the delivery of next-generation services. Already, various types of use cases for multi-UAV networks are being actively investigated. For instance, the studies in [1] and [2] have discussed the possibility of UAVs being deployed as aerial base stations, while other studies have investigated the application of UAVs for geographical surveying [3], UAV-aided relaying [4], and vehicular communications [5], [6]. Clearly, the application of multi-UAV networks has seen keen interest from both industry and academia. Compared to single-UAV networks, large-scale deployment of UAVs, i.e., multi-UAV networks, enables greater link redundancy [7], while overcoming potential weight and flying time restrictions of a single UAV [3].

Although multi-UAV networks can unlock many potential benefits for next-generation networks, supporting large-scale deployments of UAVs require readily available spectrum for UAV communications. Operations related to UAV control and non-payload communication (CNPC) links, for instance, have been allocated the L-band (0.9 - 1.2 GHz) and C-band (5.03 - 5.091 GHz) by the International Telecommunications Union (ITU) [8]. However, it is noted that the L-band spectrum is highly congested. Taking the USA as an example, the L-band is shared by terrestrial systems (0.9 - 0.96 GHz), aeronautical communication systems (0.96 - 1.164 GHz), and satellite communication systems (1.164 - 1.215 GHz), in addition to UAV CNPC systems [9].

In Singapore, CNPC and non-CNPC links for UAV operations can only operate in the 433.05 - 434.79 MHz band, 2.4 - 2.4835 GHz band, and the 5.725 - 5.850 GHz band [10]. Similar to the USA, the allocated bands for UAV communications are also congested. Maritime mobile systems, for instance, operate in the 433.05 - 434.79 MHz band, while systems with Bluetooth or wireless local area network (WLAN), i.e., WiFi, capabilities operate in the 2.4 - 2.4835 GHz band [11], [12]. It is also noted that WLAN devices also operate in the 5.725 - 5.850 GHz band [11], [12].

Given the highly congested spectrum that has been allocated for UAV communications, the potential benefits of multi-UAV networks may become negated. With limited availability of spectrum, large-scale deployment of UAVs in multi-UAV networks may not be possible. Furthermore, interference from other wireless systems sharing the same band can cause unnecessary interference to multi-UAV networks, which in turn jeopardizes the reliability of UAV communications. Therefore, a lack of available spectrum dedicated to supporting UAV communications, is itself, one of the significant challenges that must be addressed soon.

In this context, a hybrid-duplex (HBD) UAV communication system (UCS), i.e., HBD-UCS, can be explored as a viable alternative to overcome spectrum scarcity in UAV communications. In an HBD-UCS, uplink and downlink UAVs, equipped with conventional half-duplex (HD) transceivers, concurrently communicate with full-duplex (FD) ground stations (GSs) in the multi-UAV network (Fig. 1).

In this way, an HBD-UCS effectively improves spectrum utilization by doing away with the need for separate uplink and downlink bands. On the same note, it is worth pointing out that equipping UAVs with FD transceivers, i.e., FD-UCS with only FD-enabled nodes, can enable spectrum efficiency to be further boosted. Such a setup has been proposed in [13], where UAVs equipped with FD transceivers function as aerial base stations. While equipping UAVs with FD transceivers can greatly enhance spectrum efficiency, getting such FD-UCSs to be regulatory-compliant can be expensive in terms of time and resources. To see this, it is worth noting that uplink and downlink interference are simultaneously experienced at all nodes in an FD-UCS. The resultant interference is a major stumbling block towards having regulatory-compliant FD-UCSs in Singapore, as UCSs are required to share the same spectrum with other wireless systems without causing interference [10]. Therefore, extensive time and resources will be required in order to determine the impact of uplink and downlink interference from FD-UCSs on other wireless systems operating on the same spectrum.

Apart from interference, the transition towards an FD-UCS may be financially costly, as current HD transceivers available for HD-UCSs cannot be employed in FD-UCSs. As a result, the process of designing and implementing regulatory-compliant FD transceivers for UAVs and GSs can potentially increase the development cost and deployment time of FD-UCSs. In contrast, HBD-UCSs experience less interference as conventional HD transceivers can still be used for UAVs. Furthermore, only changes, e.g., new FD transceivers, at the GSs are required in an HBD-UCS, as compared to an FD-UCS which requires changes at the UAVs and GSs. Constraints on the size, weight, and power (SWAP) of UAVs may also result in FD transceivers that are either infeasible to design or non-compliant with regulatory requirements. Therefore, overcoming spectrum scarcity with an HBD-UCS allows HD transceivers on UAVs to be used, while allowing multi-UAV networks a smoother transition from HD-UCSs to HBD-UCSs.

Despite the associated advantages, self-interference (SI), stemming from FD transmissions at the GSs, and uplink interference at the downlink UAVs, i.e., inter-UAV interference, are the main impediments in HBD-UCSs [14], [15], [16], [17], as depicted in Fig. 1. Although SI can be suppressed, either in the passive domain by introducing path loss, or in the analog or digital domain via interference cancellation, residual SI is unavoidable. In particular, SI mitigation at FD-enabled GSs is imperfect due to non-ideal characteristics in practical FD transceivers, such as carrier phase noise and imperfect SI channel estimation [14], [15], [16], [17], [18]. Therefore, an important step in enabling practical HBD-UCSs starts with SI mitigation architectures that minimize the effect of non-ideal characteristics in FD transceivers. Likewise, practical HBD-UCSs require effective strategies at the downlink UAVs to manage inter-UAV interference.

Given the limitations of HBD UAV communications, accurate and realistic modeling of an HBD-UCS is required before any potential solutions can be explored. In this aspect, several surveys that discuss various aspects of UAV communications are available in the literature. In [19], several aspects of UAV communications were surveyed, including the characterization of UAV network types, routing protocol requirements in multi-UAV networks, handoff schemes in multi-UAV networks, and energy efficiency methods in UAV communications at the physical, data link, and network layers. In [20], a survey discussing the delivery of Internet of Things (IoT) services with UAVs was noted. Specifically, the authors discussed several use cases and architectures for UAV-based IoT service delivery, UAV-related regulatory requirements, along with the associated technical challenges. In a survey paper presented by Hayat et al. [21], the characteristics, applications, and requirements of multi-UAV networks were discussed from a communications perspective. Several candidate technologies that can potentially support UAV communications were also surveyed, including Bluetooth and WiFi, along with open research problems and challenges. Apart from these survey papers, similar works that surveyed UAV communications for integrated satellite-aerial-terrestrial networks [22], [23], UAV-assisted wireless networks [24], UAV-based fifth-generation (5G) and beyond networks [25], and UAV-assisted cellular networks [26] have been noted. Recent survey papers on channel modeling in UAV communications have also been seen [27], [28]. From the above discussions, a summary of the related survey papers on UAV communications is provided in Table 1.

While the above survey papers have discussed many different aspects of UAV communications, the issue of spectrum scarcity has mainly been ignored. To this end, the main objective of this survey paper is to discuss the associated challenges and opportunities in HBD UAV communications. To the best of our knowledge, the current survey paper is the first to specifically focus on the application of FD-based networks to improve spectrum efficiency in UAV communications.

In the remainder of this work, the organization of the paper is as follows (Fig. 2). In Section 2, the state-of-the-art related to UAV channel modeling is discussed, along with open research problems. In Section 3, SI mitigation architectures that can be used in HBD-UCSs are presented. Furthermore, various types of FD transceiver imperfections, along with accurate modeling of the impairments, are discussed. After that, an example signal model at the FD-enabled GS that accounts for the discussed FD transceiver impairments is presented as a case study. Discussions on key research challenges of HBD-UCSs that are regulatory compliant, and the technical feasibility of FD-enabled UAVs, are also presented. Interference management strategies that can be adopted at the HD UAVs are discussed in Section 4. Specifically, the state-of-the-art and the advantages and disadvantages associated with the discussed interference management strategies are surveyed in Section 4. Discussions on the practical implementation, performance evaluation, and open research problems, of the associated interference management strategies in an HBD-UCS are also presented. In Section 5, non-orthogonal multiple access (NOMA) methods are discussed to enable HBD-UCSs to support multi-UAV scenarios. An example signal model is then proposed for a power-domain NOMA-aided HBD-UCS, with open research problems presented for successive interference cancellation (SIC) detection, user-pairing, and DL-based techniques in HBD UAV communications. Finally, the survey paper is concluded in Section 6.