Scalable Full Duplex Dense Wireless Networks (SENSE)

Lead Research Organisation: King's College London
Department Name: Informatics


In response to the growing demands for delivery of content-rich and delay-sensitive services, network architectures for 5th generation and beyond wireless communication systems are becoming more and more dense. This illustrated through the ever increasing deployment of small cell networks as well as machine-to-machine (M2M) communications. This trend, whilst improving network capacity, will still necessitate reuse of available resources such as frequency spectrum within smaller areas by larger number of nodes/cells, which in turn would adversely affect the quality of service.

On the other hand, by allowing simultaneous transmission and reception in the same frequency band, In-band Full-Duplex Communication (IFDC) technology potentially enhances the spectral efficiency of a single point-to-point (P2P) channel by 100% over the conventional half-duplex communication. IFDC also enables the nodes, e.g. in P2P scenarios, to receive channel feedback or sense other channels whilst transmitting data, which shortens the latency compared to conventional half duplex communication with time-division-duplexing. Moreover, using full duplex relay nodes in multi-hop scenarios can potentially reduce the end-to-end latency by enabling simultaneous receiving and relaying. Practical implementation of this technology requires rigorous interference cancellation methods at each node to suppress the strong self-interference imposed on the receiver by the transmitter of the same node. The major bulk of research on IFDC has focused on self interference cancellation (SIC), and the respective state-of-the-art technology can achieve a high level of SIC at full duplex terminals; hence the IFDC technology has become closer to commercial deployment by industry.

Deploying IFDC in realistic dense settings entails new range of technical challenges, and opportunities alike. IFDC can yield substantially greater network throughputs and delay reductions over half duplex networking by deploying the technology in denser networks. However, attaining such gains demands for efficient scalable resource allocation and multi-node interference control methods. This great potential of 'full-duplex dense networks' in 'scalable service provisioning' has not been addressed to date by the research community in sufficient depth.
At physical-layer, new resource allocation challenges arise in IFDC networks; for instance, in the design of concurrent channel sensing and data transmission, and in adapting transmit power of the nodes to their variable self-interference. Also, using IFDC in dense scenarios will affect design of the protocols in the higher layers; for instance IFDC would entail greater chance of packet collisions and multi-node interference, which demands for new medium access control (MAC) protocols suited to the emerging dense full duplex networks. Furthermore, IFDC will enable full duplex relaying in multi-hop communication, hence requires new Forwarding-layer/Network-layer protocols to deal with the new full-duplex forwarding paradigms.

For conventional half duplex scenarios it is known that network throughput and quality of services can be improved through cross-layer methods, particularly with co-design of physical and MAC layers or MAC and Network/Forwarding layers. In fact for optimal scalability of heterogeneous services in full duplex dense networks, cross-layer approaches are inevitable. This project aims to propose systematic design of resource allocation and interference suppression techniques and algorithms at physical, MAC and Forwarding layers in order to enable substantial throughput gain and delay reduction by deploying full-duplex communication in dense wireless networks. These new methods will pave the way for deploying scalable service provisioning in the emerging dense wireless networks.

Planned Impact

Through the adoption of SENSE based technology for future wireless access systems, the general public will benefit from this EPSRC funded activity, given our vision of improved spectrum and power efficiencies as well as reduced network latencies. The project outcomes will have a profound impact on addressing wireless capacity predictions and end-to-end service delay for networks of 2025 and beyond, thus helping to meet public expectations of future wireless connectivity. The role of industry, and in particular our industrial partners, will be paramount here given the significance of international standards in the wide spread adoption of new wireless methodologies in the sector. The dissemination of our research outputs to the standardisation bodies (3GPPP and the ITU) as well as industry fora (MWC, Cambridge Wireless, etc) is seen as an important contribution.

Clearly, our industrial partners being closely aligned with our research, are well positioned to exploit the research outcomes within their products and services. This potentially encompasses transceiver chip set development, wireless network architecture and management, consumer products with enabled wireless connectivity as well as test and measurement.


10 25 50
Description We have devised new methods for communication between vehicles and between vehicles and road side units. This method is based on full duplex communication, a 5G technology, and will help vehicles to sense the other vehicles and road side units whilst transmitting data to others. This will improve traffic congestion and also can help in avoiding car accidents. Using full duplex technology we have produced new algorithms for supporting V2X communication of emergency information more efficiently.
We have also developed new results on full duplex relaying in UAV-based wireless networks which can be used in low-latency 5G systems. In the first half of the project we have learnt that a major use of full duplex technology will be in reducing end-to-end delay, which will be elaborated on during the second half.
During the 2nd/3rd years of the project we investigated the benefits of full duplex networking in the contexts of cognitive networks and introduced new techniques for improving throughput (capacity) and reduce delay in such networks. We specifically incorporated neural network based learning methods to improve throughput/quality of full duplex communications in cognitive networks; networks in which the nodes are more intelligent in processing the data and in learning the environment.
Furthermore, we have produced new methods for joint beamforming design at the transmitter and receiver (where multiple antennas are incorporated by the transceivers) for improving (transmission) capacity of ad hoc networks with full duplex and half duplex transceiver nodes. This method helps increasing the number of users that can communicate simultaneously whilst keeping the level of multiuser interference low.
Another finding of this project has been in relation with joint edge computing and energy harvesting in non-orthogonal multiple access scenarios (NOMA), using full duplex capability of the edge base stations. In other words, one group of users simultaneously offload task data to the base station (BS) and the BS simultaneously receive data and broadcast energy to other group of users using Full Duplex technology.
A number of PhD and MSc students were trained by the RAs of this project, and each one is already working on tackling the challenges of the technology in realistic scenarios.
Exploitation Route Through the project's industrial partners, and also through King's partners on 5G, especially Ericsson and BT. King's is in close collaboration with them in its 5G lab, and two of the investigators of this project are also part of the King's 5G lab team, which will be pivotal for findings of the project be taken forward and put to use by wider industry. We also present (and have presented already) the results of the project in major conferences of the field as well as publishing in the leading journals.
Sectors Digital/Communication/Information Technologies (including Software),Energy,Transport

Description The findings of the SENSE project were presented to industrial and academic audience in June 2018 at a workshop at King's dedicated to this project. Presentations from King's College London and University of Bristol showed the early results and roadmap of the project, and the capabilities of the underlying technologies in improving 5G networks. The project lead (Prof M. Shikh-Bahaei) also chaired a panel session to discuss in detail with the external experts about the practical challenges in implementing full duplex technology in future dense networks, and pathways to impact of this project. Moreover, the results of this work have been presented in an industrial event (IEEE CQR workshop) for benefit of industrialists in wireless com field. Also the project lead has been working with project partners and also with external industrial collaborators on addressing the challenges of implementing the technology in real life networks. We have closely worked with Ericsson, a King's partner in 5G testbed, in this respect. A major impact of the research in the Sense project has materialised in collaboration with partners of a EC Horizon 2020 project, 5GCAR. Our results on full duplex V2X communication have been published in the final EC project reports, which can potentially improve quality of autonomous cars -e.g. in crash avoidance- in the future.
First Year Of Impact 2019
Sector Digital/Communication/Information Technologies (including Software),Transport
Impact Types Societal,Economic,Policy & public services

Description SENSE 
Organisation Thales Group
Department Thales Research & Technology (Uk) Ltd
Country United Kingdom 
Sector Private 
PI Contribution We feed back to the partners the results of our research
Collaborator Contribution They attend our technical meetings and advise on the way forward and on producing impact
Impact Our research articles are partly results of this collaboration.
Start Year 2017