Research

Communications, Sensing and Imaging

The Communications, Sensing and Imaging (CSI) group at the University of Glasgow covers a wide range of both basic and applied research. We focus on first principles of engineering science with a goal to realise efficient and optimized wireless systems and components for communications, sensing and imaging applications.

Our research has been funded by major research councils and local & international funding bodies, including the Scottish Government, EU-Horizon 2020, EPSRC, CDE/DSTL, UKRI, Innovate UK, IAA, GCRF, QNRF and UAEF. Our highlight project on Scotland 5G Centre brings together academia and industry to test, develop and deploy their 5G applications and solutions on our campus wide 5G laboratory.

Communication systems research covers the majority of physical and MAC layer of wireless communications and networking with particular focus on beyond fifth generation (B5G) communication technologies. Over the coming years, 5G development and subsequent deployment will support super-fast, ultra-reliable, ubiquitous communications across large numbers and types of devices, connecting people and machines in a way often referred to as the ‘Internet of Things’ and the ‘Internet of Skills’. Our group conducts research on wide areas of 5G technology, ranging from:

1. Advanced Sensors, Antennas and Devices
2. Converged Networks and Network Resilience
3. Self-Organising Networks (SON)
4. Blockchain Wireless Networks
5. Intelligent Reflective Surfaces (IRS)
6. New Antenna Design (esp. mmWave and THz.)
7. AI-Driven Design
8. Signal Processing for Future Waveforms
9. Wireless for Healthcare
10. Internet of Things
11. Emergency Networking (popup networks)
12. Free-space Optics and Wireless Backhaul for beyond 5G networks
13. Use cases of 5G technology (For example: Smart Transport, Smart Agriculture, Connected Health, body area networks, Environment monitoring, Connected buildings, etc.)

Research topics

Network resilience
Self-organised networks (SONs)
Internet of things
Emergency networking (Pop-up networks)
Wireless healthcare
Intelligent reflecting surfaces
Free-space optics and wireless backhaul
AI-driven design
Blockchain wireless networks
Advanced sensors and devices
Signal processing for future waveforms
New antennas (THz and mmWave)
Wireless channel characterization and modelling
Physical layer of wireless/cellular communications
Energy and spectrum efficient cellular communications
Vertical markets for wireless communications
Nano-communications
Energy harvesting communications
Ultra-reliable low latency communications (URLLC)
Future Radar Systems

Network resilience

A single point of failure is a major concern in most centralised networks and its detrimental effect is even more pronounced in emergency scenarios. Hence, this research work will focus on proposing distributed data delivery and resource management schemes for improving network resilience, especially in emergency situations. Stochastic geometry will be used to model nodes and the location of network entities along with appropriate traffic models to develop theoretical, as well as practical understanding of the system. The performance of these algorithms will be evaluated in terms of carefully selected key performance indicators.

Self-organised networks (SONs)

5G & beyond networks will rely on ultra-densification to achieve and maintain targeted key performance indicators (KPIs). An extreme variety in services and corresponding KPIs are expected in future networks. AI-enabled self-organised networks (SONs) can guarantee energy efficiency, low-cost operation and ubiquitous coverage, and it can make the network proactive rather than reactive. In this project, we will develop AI-enabled proactive SON functions on load balancing, resource optimisation, mobility and handover management, and energy saving. The project will also design agile and scalable self-healing functionalities for ultra-dense future cellular networks.

Internet of things

The advent of Internet of Things (IoT) will lead to billions of new connected devices which will place a strain on today’s cellular and WiFi spectrum. Terahertz communications can revolutionise wireless networks while addressing the spectrum shortage through the massive high frequency spectrum. Hence the technologies of both the next generation cellular networks and IoT will move toward some form of convergence, where the use of terahertz frequencies will become commonplace in addition to the conventional cellular band. This project will develop AI enabled solutions to deliver seamless connectivity experience to the user to manage the emerging convergence. This project also develops efficient solutions and algorithms on beam search procedure in terahertz, AI enabled mesh-based topologies for terahertz communications. The project will also address the seamless switching between WIFI, cellular, and terahertz technologies while considering novel architectures such as the multi-tier control and data plane separated architecture.

Emergency networking (Pop-up networks)

Disasters are a huge threat to communication systems either due to network infrastructure being destroyed or the network being overwhelmed by a sharp rise in traffic, thereby degrading network performance. Current systems are expensive to operate and maintain in times of emergency, and require significant human intervention, resulting in long commissioning and decommissioning times. This work aims to design a pop-up emergency network, based on the principles of self-organisation (configuration, optimisation and healing), that takes into account the peculiarities of emergency communications – power, backhaul and deployment time. The use of entities such as small cells and Unmanned Aerial Vehicles will also be explored to address the challenges of emergency communications.

Wireless healthcare

The healthcare paradigm is shifting from doctor-centric to patient-centric concept. The prevailing techniques to enable this paradigm shift uses ambient sensors, cameras and wearable devices that primarily require strenuous deployment overheads and raise privacy concerns as well. To overcome aforementioned issues, an emerging technique to use perturbations in Channel State Information (CSI) of wireless signals for detecting human movements (which are linked to various health conditions) is attracting attention due to its non-invasive nature and security feature. Our work is focused on the development of non-invasive, easily-deployable, flexible and scalable testbed for identifying large-scale and small-scale body movements based on Software Defined Radios (SDRs).  By employing, machine learning on the collected data, we are able to do proactive health measures.

Activity Recognition and vital signs monitoring: healthcare and veterinary applications

Research applications include the analysis of human and animal radar signatures across different domains of data representations for classification and vital sign monitoring. This can be applied to classify different activities (walking, running, human interactions), as well as healthcare (e.g. fall detection, activities of daily living, and out-patient follow up care) and lameness assessment for precision livestock farming (cattle, sheep) and leisure/racehorses. These works encompass signal processing, machine learning, and multimodal sensing including radar, accelerometers, IMUs.

Intelligent reflecting surfaces

Intelligent reflecting surface (IRS) is a novel concept that is been introduced in beyond 5G communication. It will reconfigure the wireless propagation environment via software-control reflection. This surface will consist of low-cost integrated electronics that would sense and reflect electromagnetic waves in a specific direction. As a result, this would boost the signal at the receiver with no added hardware cost. The aim of this research work is to develop and model a hardware testbed for the IRS. The proposed work will explore various meta-surface unit cells which could effectively manipulate the electromagnetic waves and design a control board with IRS for beamforming applications. The hardware imperfection and energy efficiency will be considered as main metrics to optimize.

This project will investigate the optimal IRS beamforming and beam management algorithms, both analytical and simulation results will be provided to guide the practical IRS deployment.

Free-space optics and wireless backhaul

As we move towards 5G and beyond, the need for overcoming the issue of spectrum-crunch is testing the nerves of reliability and security of our modern-day communication systems. This can be efficiently tackled via utilising non-traditional spectrum range that is license-free in its availability and inherently secure while being enormously robust. FSO technology boasts of being complimentary to the traditional radio-frequency (RF) communication systems and millimetre wave (mmWave) technology, among others. Hence, based on this highlighting feature, FSO technology will be the source solution towards wireless backhaul. Specifically, we address issues in modelling FSO wireless transmissions and analysing its performance while this technology is utilised with RF and mmWave technologies in complementary fashion. Based on this analysis, we have been deeply investigating these systems from physical-layer (PHY) security perspective as well. Subsequently, we have devised adaptive algorithms in improving throughput of mixed RF-FSO systems.

AI-driven design

Integrated circuits, antennas, arrays, filters, and multiplexers are essential in communication systems. Present manual design methods suffer from long time-to-market and often suboptimal design quality. To address this challenge, we propose novel AI (evolutionary computation and machine learning) techniques to assist or automate the design. Our AI-driven antenna design methods, the SADEA series, can obtain highly optimal designs with up to 20 times efficiency speed improvement compared to other existing methods. Moreover, the initial design is not needed. The SADEA series ranks first in comparisons carried out by MathWorks and Altair and was embedded into MATLAB. In particular, we invented the first AI-driven design method for 5G base station antennas (sample here). We also invented the first method for the automated filter, diplexer, and analogue IC automated design ready for industry use by addressing key bottlenecks. They are essential tools for more than 10 leading design teams (both academia and industry). We believe that induced significant improvement in design quality and time-to-market are also applicable to BT.

Blockchain wireless networks

Blockchain is primarily designed in stable wired communication environment running on advanced devices, which is not suitable for high dynamic fading wireless connected digital society that composed of massive low-cost Internet of Things (IoT) devices. Wireless blockchain networks will analyse and optimise the blockchain security and performance such as transaction throughput, latency, and scalability, through jointly designing blockchain and wireless communication system architecture, protocol, algorithms and deployment.

Advanced sensors and devices

Tele - Point of Care (PoC) tools for food safety, agriculture/aquaculture, healthcare (prognostic and diagnostic). Onsite testing via mobile phone apps and data transfer for offsite analytics.

Signal processing for future waveforms

Performance estimation for the fusion of radar and communication signals from modelling to experimentation. With further device integration and demands to offer more services for customers or gather more information for mobile radio head intelligence, the integration of radar to fulfil both the communication and radar function simultaneously, however the choice of waveform and architecture is specific to the usage and the designed performance metrics. CSI can propose a study for task technology fit in order to design a waveform suited for the application and estimate performances and hardware requirements to reach set goals.

New antennas (THz and mmWave)

Demand of high transmission data rates, low latency, high reliability and interference-free operation for applications in communications, infotainment, autonomous vehicles, positioning & localisation, smart healthcare and smart agriculture is ever-increasing. The need is driving the development of wireless networks fostering the exploration of new spectrum along with the use of conventional frequencies and Millimeter-wave (30 to 300 GHz) and Terahertz (0.1 to 10 THz) bands are considered as the front-runner enabling technologies for future wireless networks. High performing wireless devices require efficient low profile antennas offering larger bandwidth, higher gain and insensitivity to the human user presence while maintaining reasonable performance in ever-shrinking form-factors and under extreme interference conditions to ensure reliable communications.

The antenna design cluster in CSI is taking on this challenging task and provides insight into the antenna design considerations fostering novel approaches for the development of efficient, cost-effective, scalable, and reliable antenna solutions for wireless networks and Internet of Things. Some of the key areas being investigated are massive MIMO antenna systems and beamforming techniques, smart reconfigurable and multiband antennas, antennas for wearable and implantable devices, base station and mobile terminal antennas, antennas for Machine-to-Machine (M2M) communications, phased array antennas, RFID antennas, GNSS antennas and AI-enabled antenna solutions.

Wireless channel characterization and modelling

Focus is on parametric stochastic channel modeling of small scale and large scale variations in wireless propagation channels. Contrary to other approaches widely employed, such as measurement/site specific-, simulation (e.g., FDTD, ray tracing)-, or geometry-based that heavily rely on the specific wireless scenario setting (measurement or site specific-based) or fail to accurately model all features of multipath propagation such as multi-bounce scattering (geometry-based), we have been adopting a parametric stochastic modeling approach that can adapt to every propagation scenario, e.g., urban cellular, vehicle-vehicle, point-to-point mmwave communications. The inherent spatial, temporal and frequency variations are stochastically modelled following a duality principle, i.e., concepts from spatial variations modeling to be adopted in frequency variations modeling, etc. The applicability and limitations of all channel modeling approaches is comprehensively considered in terms of accuracy and complexity. One more advantage is that the adopted methodology can be readily adapted in the design of a wide range of wireless components and systems such as, multi-antenna systems, OFDM, CDMA receivers and physical layer security modules.

Energy and spectrum efficient cellular communications

Our research focuses on reducing the energy bills for the network operators as well as saving the planet by reducing the carbon foot print of cellular networks. We look at energy efficient design at component level, energy harvesting solutions, efficient design of node level subsystems including power amplifier and then system level solutions like discontinuous transmissions and efficient cell-muting. Cognitive radio networking is further exploited for high spectrum efficiency via cooperative and cluster based spectrum sensing algorithms. With the objective of minimizing leased spectrum cost, the user requests are served with the sensed spectrum or put in a time bound queue in case of free spectrum unavailability. With the objective of minimizing leased spectrum cost, the user requests are served with the sensed spectrum or put in a time bound queue in case of free spectrum unavailability. We work on fundamental performance limits as well as practical solutions approaching these limits. We also look at disruptive new technologies like Device to Device and mm-Wave for their potential of energy efficient communications. We have pioneering publications on the futuristic architecture of Control Data Split (CDSA) to enable energy efficient operation of cellular systems.

Vertical markets for wireless communications

As a research group, we holistically exploit our capabilities to provide solutions to the vertical industries exploiting the capabilities of wireless communications. This includes (but is not limited to) smart cars, smart metering, smart manufacturing, financial markets, ultra-reliable communication, tactile internet, cost-effective solutions for covering rural areas, mHealth, Healthcare for underprivileged and Health Informatics, communications for learning and teaching solutions, delay tolerant services, use of Unmanned Aircraft Vehicles UAVs for different services, communication for reshaping energy demand, smart grid, Smart spaces, Mobile phone applications and many more areas.

Nano-communications

With the development of the nanotechnology, the idea of coordinating the nano-devices was put forward, leading to the appearance of the nano-network.  Nano-technology has a critical role now a days in multidisciplinary domains such as environmental, industrial, biomedical and military; one of the emerging social and scientific impact of such technology would be in healthcare and bioengineering applications.  The main objective of this research is to investigate the possibility of connecting nano-devices securely by using various communication paradigm for different applications like biomedical and environmental. Emphasis of research will be on developing novel channel models, investigation of feasible modulation techniques and new physical layer security protocols for nano-devices.

Ultra-reliable low latency communications (URLLC)

5G systems simultaneously support various services (use cases) and need to span a wide range of requirements. Besides the race for increased data rates in 5G networks, enabling ultra-reliable low latency communication (URLLC) is another challenge in 5G networks. Notion of reliability in URLLC not only includes reliable transmission of data in a network, but it includes time dimension as well. The data has to be delivered within the latency targets reliably and the latency target are for end-to-end system. URLLC finds applications in health sector, autonomous vehicles, industrial control, mission critical applications, and many other emerging areas. The challenge is to develop resource allocation mechanisms in a wireless network that enable URLLC. The topic of interest in this emerging area of research include techniques like wireless proactive caching, fog computing, cloud RAN, hybrid ARQ, Device to Device communication and network slicing to name a few.

Future Radar Systems

With the growing overcrowding of the spectrum, new threats (e.g. UAVs near airports) and the need for spectrum reuse, there is a need in developing novel radar systems that can evolve in that environment and communicate amongst themselves. This starts with developing the radar spectrum sensing and actively communicating with neighbouring radars or communication nodes to collaborate and fuse information. These functions have to be executed alongside radar functions and switch from one mode to another or combine several modes together in its operation via resource management. Having a software-defined radar platform to do this would mean that the radar could dynamically change its operating mode to fit the task at hand by acting via software on both hardware and processing configurations. We explore waveform design, MIMO, radar geometries, and polarisation to that end.