2017 Projects Awarded in the first Call for EDGE Proposals

Flash Report on the results of the first call – published 6th April 2017

WriteAid: Machine Translation-Assisted Medical Publication

This project aims to provide writing assistance to professionals working in specialised domains through machine translation (MT). In so doing, the project will facilitate better and broader communication of scientific content globally. Researchers and professionals working in specialised domains such as the medical sector, need to publish their work in scientific conferences and journals nowadays and face the challenge of publishing in English to increase the outreach and impact of their activities. Although English is the de facto lingua franca for scientific research (Bennett, 2013), relatively few researchers are native speakers of English. This leads to the need for mastering a second language or hiring translators or professional revisers to ensure that professional content production meets the language quality required. In fact, more revisions are usually required for papers written by English as a Foreign Language (EFL) writers (Lillis and Curry, 2010. The idea behind WriteAid is to explore the use of MT to assist researchers to write scientific papers in English by allowing them to use their mother tongue to write them and subsequently asking them to post-edit a machine translated version of their paper. The results of this project will allow medical researchers to publish papers in English without cumbersome technical processes. Moreover, such results could be extrapolated to other language pairs and domains, thus helping researchers wishing to publish their work in English, as well as Publishing Companies, alleviating the editing time prior to publishing.

  • Project Name: WriteAid: Machine Translation-Assisted Medical Publication
  • Researcher: Carla Parra Escartín
  • Supervisor: Sharon O’Brien
  • Institution: DCU (ADAPT Centre)
  • Duration: 24 Months

BAIT: Bilingual Association in Neural Machine Translation

Automatic translation of text from one language into another is a major scientific challenge. Phrase-based approaches that tile together translation fragments into complete translations have become highly successful and are currently also the basis of most commercial systems. But phrase-based translation generalizes poorly and makes strong independence assumptions in modeling the statistical translation process. Therefore, recently neural machine translation (MT) that treats translation as a sequence-to- sequence prediction task was proposed as a viable alternative. Neural MT overcomes many of the weaknesses of phrase-based translation, importantly enabling much weaker independence assumptions, and has consequently started to dominate MT research. But current neural MT has two big weaknesses. First, it provides no strong model of what words to translate together. Second, it forms no explicit representations of deliberately chosen subsets of the input words. I propose to tackle these weaknesses in two steps. First, I will add a model component that learns to make joint predictions about what source words to translate together: a bilingual association model. I will investigate the best way to train this type of model. I will also research how to best use different types of information for training and prediction, using not only the source and target words, but also syntactic information and external word alignments. Second, I will create representations for subsets of input words predicted to be relevant by the bilingual association model. Such source phrases can be discontiguous. Then I will use these representations together with the bilingual association model and representations for single words to produce better translations in neural MT. The research will give new insights in how to create effective models for neural MT that overcome its current weaknesses, and is expected to improve the state-of- the art in this high impact research field.

NoCorrosion: Computational design of non-corrosive metallic alloys

The NoCorrosion project aims at developing a methodology for designing corrosion-resistant materials through computational high-throughput approaches. In particular, the heart of the method consists in developing simple ‘descriptors’ for the chemical process of interest (surface reactivity and saldability on metal supports in the case of corrosion-resistant materials). These will be used to scan materials databases, so as to single out the most promising alloys. ‘Descriptors’ can be defined as empirical relationships between a microscopic quantity of the material (typically structural or electronic features) and the property of interest. The combination of the experience of the applicant in chemical bonding and electronic structure analysis with the well-established expertise in high-throughput materials design of the supervisor is expected to fruitfully generate a physically sound methodology that can be later exploited by the whole materials science community. Another strength of the proposed project is the direct interface with Nokia Bell Labs that are interested in the development of corrosion-resistant, environmentally friendly (namely Pb-free) solders for communication technology. The double-way feedback between the applicant and Nokia Bell Labs will create a synergy that culminates in the design of new corrosion-resistant soldering alloys.

  • Project Name: Computational design of non-corrosive metallic alloys
  • Researcher: Gabriele Saleh
  • Supervisor: Stefano Sanvito
  • Institution: Trinity College (Amber)
  • Duration: 24 Months

DEFINITE: Development of a Light Fidelity (LiFi) demonstrator based on the use of a 3D Nanoscribe printer for Telecommunications applications

This project aims to develop a novel Light-Fidelity (LiFi) demonstrator with components manufactured using a 3D Nanoscribe printer with applications in optical wireless communications. The technique to be used by the 3D Nanoscribe printer is direct laser writing which is based on two-photon polymerization (TPP), allowing for the fabrication of arbitrary 3D structures in suitable photoresists with submicron resolution. Using a Nanoscribe 3D printer we will prototype optical components such as waveguides, lenses as well as hybrid optoelectronic components for a transmitter and receiver front-ends of a LiFi system. This research will demonstrate impact by presenting innovative ways to design components in a faster and simpler way allowing for rapid prototyping, and improving the performance of devices in terms of speed and bandwidth in order to build the LiFi system. The functionality of the LiFi system will be verified at the CONNECT centre in Trinity College Dublin targeting data transmission at 100 MB/s or better. The development of materials technology in 3D photonic manufacturing within the AMBER centre is a great step towards building photonic integrated components for the future communication networks.

  • Project Name: Development of a Light Fidelity (LiFi) demonstrator based on the use of a 3D Nanoscribe printer for Telecommunications applications
  • Researcher: Josue Parra Cetina
  • Supervisor: John Donegan
  • Institution: Trinity College (CONNECT)
  • Duration: 24 Months

PhySim11p: A Realistic Physical Layer Simulation Model for IEEE 802.11p

Future cars will employ wireless communication to exchange information directly with each other and maybe infrastructure nodes, forming a Vehicular Ad Hoc Network (VANET). Through communication, cars will be able to drive cooperatively, which will make transportation safer, more efficient, and more comfortable then ever before. Today, where regulatory bodies are pushing the technology forward, industry and academia are working on applications for day one. The success of these applications will be crucial, especially for the general acceptance of the technology. To design and evaluate VANET applications, researchers rely heavily on network simulations, often using one of the well established Open Source tools like ns-3 or Veins. While both include detailed models for mobility, signal shadowing, and communication protocols, they fall short with regard to the physical layer. Especially in VANETs, the high mobility, short wave lengths, and large interference domains, lead to characteristic physical layer effects that have to be considered in great detail. Currently, the impact of these physical layer effects is well recognized by the communications engineering community, but not adequately reflected with state-of-the-art network simulators. In this project, we address this issue by developing a realistic simulation model for the physical layer. Building on results from the communications engineering community, we will use Software Defined Radio (SDR) to characterize physical layer performance and derive a computationally efficient and easy to use simulation model. Using SDR, we make results from physical layer studies accessible to the networking community, bridging the gap between Electrical Engineering and Computer Science. Our model will improve the accuracy and realism of simulations, allowing us to better assess the performance of protocols and, ultimately, to make more informed decisions regarding the technologies that will be deployed in tomorrows cars.

  • Project Name: A Realistic Physical Layer Simulation Model for IEEE 802.11p
  • Researcher: Bastian Bloessl
  • Supervisor: Linda Doyle
  • Institution: Trinity College (CONNECT)
  • Duration: 24 Months

TAPIR: Temporal Aware Personalised Information Retrieval

“The ways in which humans interact, seek and utilise information, defined as information behaviour, changes over time. People’s interests change and evolve. The goals we wish to accomplish and the tasks we perform are inextricably tied to time. Unexpected events or serendipitous discoveries attract the users’ attention towards new information. When building personalised computing systems, the absence of a proper model of time in relation to the user hinders such systems from providing timely and appropriate information and services.The extraordinary proliferation of smartphones has also been changing the way humans interact with information. We are constantly surrounded by an enormous amount of information and as a result our attention is often misspent in consuming worthless information. To alleviate users from this burden, a new generation of temporal-aware user models is required to support users in the seeking and consuming of information and by recommending timely and contextually appropriate information. Meanwhile, the explosion of Big Data represents a gold mine for capturing the information behaviour of users, making the creation of temporal-aware user models feasible.

The goal of this research is to understand how temporal dynamics -time-based patterns and trends- affect information behaviour in relation to user preferences. Which factors play a major role in information seeking? How can these factors be modelled to produce the most representative user model? How can this model be exploited to deliver the best user experience? The project contribution will be a breakthrough with the current literature resulting in a life-long user model aware of these temporal dynamics. Prototypes, developed in collaboration with the industry partners, will represent a proof of concept for resulting theories and methodologies. The research will benefit from recent developments in Personalisation & Adaptation, Machine Learning and Natural Language Processing by fostering the well-established expertise provided by the ADAPT centre.”

  • Project Name: Temporal Aware Personalised Information Retrieval
  • Researcher: Annalina Caputo
  • Supervisor: Séamus Lawless
  • Institution: Trinity College (ADAPT Centre)
  • Duration: 24 Months

FASTER: Fabrication and characterisation of athermal Slotted Fabry-Pérot laser overlayed with polymer

The proposed research here is pioneering work in the development of energy efficient active devices for telecommunications. As the optical network continues to advance towards fiber to the home, the cost of the network operation must drop and this requires devices which are much more energy efficient than present generation. For lasers, this requires laser that will operate without inefficient Peltier devices. In addition, the output wavelength of lasers changes when the temperature of the device changes under pumping disrupting their performance in wavelength division multiplexing. The aim of my project is to develop an athermal slotted Fabry-Perot (FP)laser. An athermal laser is one in which the wavelength is insensitive to temperature change. These slotted single mode laser scan be realized by simply introducing gratings in the form of distributed reflective defects (slots) into the ridge of a conventional ridge waveguide FP laser. This design approach results in significantly simpler fabrication process as well as lower costs. An original idea has been used in other photonic devices is to overlay the structure with a polymer (PSQ-LH) or theoretically another material (TiO2)with a negative thermo-optic coefficient could be used, to counteract the positive thermo-optic coefficient of all III-V semiconductors. To my knowledge this technique has not been investigated experimentally in active devices such as slotted lasers. This project will make an important contribution to the development of energy efficient devices and paves the way for integration of these devices in a photonic platform for energy efficient optical-communications since the dependence of the wavelength on the temperature is reduced or removed.

  • Project Name: Fabrication and characterisation of athermal Slotted Fabry-Pérot laser overlayed with polymer (FASTER)
  • Researcher: Foued Selmi
  • Supervisor: John Donegan
  • Institution: Trinity College (CONNECT)
  • Duration: 24 Months

HST: Heat and Spin Transfer in Hyperbolic Thin-Film Metamaterials

We plan to utilize thin-film and 2D layered metamaterials for use within a Heat-Assisted-Magnetic-Recording, or HAMR, hard drive to allow for the efficient transfer of heat and magnetization from power source to recording media. The challenge faced by researchers is to heat an area smaller than 50×50 nm2 above the Curie temperature as required for state-of-the-art bit writing in magnetic hard drives. This must be done while keeping temperatures of the partially metallic waveguide, magnetic write pole, and adjacent bits as close as possible to ambient temperatures. At present, two of the largest impediments for producing a manufacturable HAMR drive is overheating of the magnetic write pole as well as the waveguide used to deliver power to the recording media. Our thin-film metamaterials are aimed at improving heatsinking of the device while reducing unwanted reflections which can potentially damage the waveguide. Moreover, we aim to design a magnetic, antireflective (AR) coating composed of thin films, which not only improves absorption efficiency but in principle is able to induce a spin-dependent Seebeck effect. This effect will enable the transfer of spin from a more durable AR coating to the recording layer of the media which can now be further displaced from the plasmonic waveguide and write pole. The increased displacement between write pole and recording layer will greatly reduce the constraints on supplying efficient heatsinking within the device. It will also reduce burnishing of the recording layer by eliminating the risk of contact with the waveguide. We plan to investigate thin films which demonstrate a near-zero permittivity due to their antireflective properties and combine them with novel magnetic, metamaterial stacks known for producing slow light and reduced dispersion. Altogether these phenomena will be able to enhance the transfer of energy while simultaneously improving the nanofocusing of incident light.

  • Project Name: Heat and Spin Transfer in Hyperbolic Thin-Film Metamaterials
  • Researcher: Frank Bello
  • Supervisor: John Donegan
  • Institution: Trinity College (Amber)
  • Duration: 24 Months

SERHENA: Service-aware Resource Orchestration in Hybrid Cloud/Edge Network Architecture

Today we witness the explosive growth of mobile data traffic and the unprecedented deployment of demanding communication and computing services that are offered by various service providers. These developments herald the advent of a heavenly landscape for the users, but on the other hand challenge the performance and even the economic viability of mobile networks. Therefore, it is not surprising that operators are actively seeking methods to accommodate this demand in a cost-effective fashion.
A very promising solution for this problem is edge-networking (EN): the deployment and smart management of network, computing and storage resources at the network edge. This approach can substantially improve the services offered to users, and reduce the required network expenditures. To this end, SERHENA will deliver a multi-faceted optimization framework aiming to unleash the full potential of edge networking by addressing certain bottleneck issues. SERHENA incorporates many innovations, as it will: optimize the resource dimensioning in hybrid EN systems that include both central and edge resources; jointly orchestrate different types of resources (e.g., computing and bandwidth) so as to satisfy the emerging types of services; study methods to leverage idle user-owned equipment; propose online resource management policies that achieve asymptotic cooperation equilibriums; focus on service-aware network optimization techniques; and analyze the market dynamics among the service providers and operators in this new era so as to create sustainable solutions.
SERHENA will follow a novel methodology, using both rigorous analytical tools, and data-driven or testbed-based evaluation approaches. It builds on the prior experience of the candidate and the hosting team, and will leverage the very pertinent and now-developing activities of top-industry labs that are associated with CONNECT.

  • Project Name: SERHENA: Service-aware Resource Orchestration in Hybrid Cloud/Edge Network Architecture
  • Researcher: Jungho Kwak
  • Supervisor: George Iosifidis
  • Institution: Trinity College (CONNECT)
  • Duration: 36 Months

SupraBioInks: Development of Advanced Supramolecular Bioinks for 3D BioPrinting in Tissue Engineering Utilising High Throughput Approach

The emergence of 3D bioprinting in tissue engineering has delivered new hope in addressing the un-met need for complex human tissues for transplantation arising due to the limited availability of donors. 3D bioprinting represents an exciting new development whereby cell-laden biomaterials can be printed layer-by-layer into clinically relevant sizes, shapes and structural integrity depending on the tissue of interest. In particular, this approach has been utilised to address the dire need for new constructs to treat musculoskeletal disorders. The main requirement is the ability to print complex high resolution tissue structures with high structural fidelity. Furthermore, these structures should facilitate cell initiated remodelling and the bioinks should be able to modulate cell phenotype within the structure. Current systems require a trade-off between these two traits. The lack of advanced bioinks that can overcome this trade-off represents the most significant obstacle to utilising 3D bioprinting in tissue engineering. To overcome these limitations a self-assembling hydrogen bonded supramolecular hydrogel is proposed that will allow the generation of a novel library of 400 hydrogels that can readily encapsulate cells and be applied to high throughput 2D and 3D polymer array platforms for the development of advanced bioink formulations. These will be used to bioprint complex 3D bone and cartilage tissue engineered constructs. The composition of these hydrogels (an acrylic acid, a peptide and a peptide nucleic acid (PNA) base component) allows wide ranging control over the material properties and how cells will interact with them. This library will be used to screen mesenchymal stem cells on 2D and 3D arrays and the advanced bioink formulations identified will be compared against commercial bioinks for complex 3D bone and cartilage tissue engineered constructs. This novel self-assembling supramolecular bioink development strategy will be applicable to a wide variety of cell types and tissue engineering applications.

  • Project Name: Development of Advanced Supramolecular Bioinks for 3D BioPrinting in Tissue Engineering Utilising High Throughput Approach
  • Researcher: Cairnan Duffy
  • Supervisor: David Hoey
  • Institution: Trinity College (Amber)
  • Duration: 24 Months

FREEMAB: Environmentally friendly carbon cathode electrocatalysts for metal-air batteries

Recent developments in battery technologies, including energy storage and utilization, have resulted in the use of energy from renewable sources remarkably expanding. However, the current Li-ion batteries still require a substantial improvement in its specific energy to reach cost-competitive electric vehicles (EVs). Metal-air batteries are considered to be the most promising technology for such applications due to their high specific energy. Nevertheless, there are still numerous technological challenges that must be solved to find widespread application and render metal-air batteries commercially feasible. One major challenge to overcome in alkaline metal-air batteries is the development of novel air cathodic materials. Ideally, new materials must exhibit a high activity towards the oxygen reduction reaction (ORR). Additionally, the material must act as bifunctional catalysts to also carry out the oxygen evolution reaction (OER) in rechargeable metal-air batteries. The aim of FREMAB is to develop new hierarchical carbon-based materials as catalysts for air electrodes in the cathode of metal-air batteries. The project aims at investigating novel materials based on carbon and non-precious metals as bifunctional electrocatalysts for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) in rechargeable alkaline metal-air batteries. The novelty of the project lies in combining a carbon-based material with different pore sizes (micro-, meso- and macropores) with the presence of non-precious metal active sites for both ORR and OER. The durability of the bifunctional air-catalyst is also expected to be improved by using these new materials. The project combines the strengths of the applicant in catalysis with new training in advanced material development, project management, communication skills and research independence. Furthermore, the 3 month secondment at CEGASA Portable Energy S.L. (CEGASA) will offer the applicant a new research direction and will enhance inter-sectoral cross-European collaboration.

  • Project Name: Environmentally friendly carbon cathode electrocatalysts for metal-air batteries
  • Researcher: Carlota Domínguez Fernández
  • Supervisor: Paula Colavita
  • Institution: Trinity College (Amber)
  • Duration: 24 Months

CONTACT: Contact Strategies for Transition Metal Dichalcogenides

Transition metal dichalcogenides (TMDs) are two-dimensional layered materials with the formula MX2, where M is a transition metal (Mo, W, Hf, etc.) and X is a chalcogen (Se, Te, S). These materials are of interest due to their inert surfaces and thickness-dependent electrical properties. TMDs have a range of applications, as silicon replacements for the channel material in ultra-scaled metal oxide semiconductor field effect transistors (FETs), photodetectors, or in spintronic devices. However, all of these applications require metal contacts with sufficiently low contact resistance, which have yet to be demonstrated.

There are two geometries possible for forming contacts on TMDs; 1) top-contacts, where the contact material is deposited on the surface of the TMD, or 2) edge contacts where there is a lateral heterojunction between the TMD layer and the contact material. In this project we propose investigating and optimising both of these contact geometries, and studying the impact on contact resistance and FET performance.

For top contacts, we will study the impact of local doping of the TMD channel under the contact to reduce the Schottky barrier height and depletion region width, enhancing the tunneling from metal to TMD. This will be achieved by using plasma irradiation to first generate chalcogen vacancies in the top TMD layer and filling these vacancies with the plasma species as substitutional dopants.

For edge contacts we propose using a similar method to that for silicide contacts in silicon and III-V materials. A reactive metal is deposited on a few layer TMD film and a subsequent thermal anneal is used to promote reaction and inter diffusion. Consistent with recent publications, for certain metals, this will result in the conversion of the underlying TMD to a metal-TMD alloy, thus forming the edge contact.

  • Project Name: Contact Strategies for Transition Metal Dichalcogenides
  •  Researcher: Lee Walsh
  • Supervisor: Paul Hurley
  • Institution: Tyndall National Institute (Amber)
  • Duration: 24 Months

DISCANT: Domain-Independent Semantic Annotation of the Text

Surrounded by huge and exponentially growing volume of information of various nature, every day we face challenges with keeping track of the latest news in a domain of interest or finding good-quality answers to specific questions. The problem of information overload has been addressed by modern information systems and search engines, however, their capabilities are strongly limited by using unstructured textual formats for exchanging and storing information. Textual documents often contain a lot of concrete facts and quantifiable information, but the communication of these facts and information through human languages renders them inaccessible to machines. This “semantic bottleneck” problem substantially slows down communication and knowledge propagation in the world. In order to equip machines with the ability of effective processing of the text, DISCANT (“Domain-Independent SemantiC ANnotation of the Text”) aims at creating a comprehensive framework for semantic annotation of textual documents of arbitrary domains, such as scientific papers, legal documents, customer reviews or clinical trial reports. We will develop approaches, methods and tools for two classes of solutions: an environment for discovering entity and relation types in a given domain and a system for automated semantic annotation of the text. The project proposes a novel approach based on a combination of unsupervised natural language processing and machine learning techniques. DISCANT will advance machine understanding of the text, contributing to the release of important knowledge buried in textual documents, the creation of machine-readable knowledge repositories and more effective solutions for semantic search and personal recommendations. As a consequence, DISCANT will equip the consumers of textual documents with better tools for overcoming data deluge and information overload, enabling them to make better-quality, data-driven decisions.

  • Project Name: Domain-Independent Semantic Annotation of the Text
  • Researcher: Dominika Tkaczyk
  • Supervisor: Joeran Beel
  • Institution: Trinity College (ADAPT)
  • Duration: 24 Months

COCODIMM: Complexity and Control of Distributed Massive MIMO

Massive MIMO in both, its co-located and distributed variant, has become an important part of future 5G plans so it needs to be understood well to be fully utilised. The nature and the scale of massive MIMO systems both imply existence of complexity observed in complex systems science manner, especially in the distributed massive MIMO case. Complexity in massive MIMO has not been investigated so far, therefore this research aims to fill this gap and grasp its complexity features by providing quantification, qualification and useful models of massive MIMO as a complex system. As a natural extension, this research also deals with devising control algorithms based on Massive MIMO models developed from complexity standpoint, verification of these algorithms and their implementation on real Massive MIMO systems, together with a thorough mathematical analysis using tools from control theory. Finally, this research will introduce reversible computing in Massive MIMO, investigating its effects on control and performance. Results of the research are diverse, ranging from publications explaining complexity in Massive MIMO setting, over novel tools and metrics developed for the specific application of complex systems science to Massive MIMO, to functional controller prototypes based on algorithms developed. This in turn represents the multidisciplinarity of the proposal as well, combining telecommunication systems with complex systems science, control theory, but also embedded system design and computer science. This proposal also includes some promising preliminary results of cellular automaton modelling and antenna selection control of Massive MIMO system.

  • Project Name: Complexity and Control of Distributed Massive MIMO
  • Researcher: Harun Siljak
  • Supervisor: Nicola Marchetti
  • Institution: Trinity College (CONNECT)
  • Duration: 36 Months

Delivery of Proangiogenic Drugs in Diabetes

Diabetes mellitus is the result of decreased insulin production or increased insulin resistance, and currently devastates the lives of millions of people worldwide through numerous complications resulting from vascular dysfunction. Pancreatic islet cell transplantation is a potentially curative therapy for treatment of diabetes mellitus, however current strategies are hindered by poor cell survival and retention. We propose to develop a clinically feasible and innovative platform for Delivery of pro angiogenic drugs in Diabetes (DAD) that promotes robust early vascularization of the transplant niche ensuring sustained cellular function and insulin production in transplanted islet cells. In addition, the DAD delivery platform will be extended to improve diabetic wound healing. These goals are supported by the following objectives:

  1.  design of novel pro-angiogenic microparticles (PAMs) to induce vascularization of islet cell implants
  2. validation of angiogenesis by PAMs in vivo
  3. design of a novel PAM-coated balloon catheter for engraftment site pre-vascularization
  4. modification of PAM therapy for treatment of diabetic wounds in vivo.

DAD will advance the state of the art by incorporating a staged approach to islet cell transplantation and combining innovative drug delivery strategies for vascularization pre- and post-implantation to generate a stromal support network for rapid engraftment and improved islet cell function. This proposal capitalizes on my unique background as a surgeon-scientist and brings together inter-disciplinary leaders in pancreatic islet transplantation, biomedical engineering, pharmaceutics, stem cell biology, endocrinology, and vascular surgery from across Europe and the USA. DAD will serve to rapidly produce clinically relevant technologies to ease the worldwide burden of Diabetes Mellitus. It will also be an ideal mechanism to re-establish my presence in research by growing me thereby advancing me toward my career goal of becoming an independently-funded leading surgeon-scientist in the area of regenerative medicine.

  • Project name: Delivery of Proangiogenic Drugs in Diabetes.
  • Researcher: Scott Robinson
  • Supervisor: Garry Duffy
  • Institution: NUIG (Amber)
  • Duration: 24 Months

Big Data Driven Geo-Spatial Models Supporting Public Health

In the emerging era of big data, digital sensing is integrated into our physical environment, creating an “Internet of Things” (IoT). Data from sensor networks, RFID tags, CCTVs, drones, and geotagged social-media posts all have geographical components. Moreover, people also often become living, breathing, walking sensors thanks to smart phones and tracking devices. All this information is fuel for geospatial modelling and predictive analytics that can provide insights which could not be discovered in other ways. Such data overflow opens new opportunities for research not possible before with old, disconnected analytics and modelling. 

This research aims to improve public health through the combined use of cutting edge technology including IoT, sensor networks, big data, and geospatial models. Based on data obtained from ground and mobile sensors, established GIS procedures, machine learning, predictive analytics and spatiotemporal modelling will be applied to achieve the following specific objectives having Dublin city as a case study area:

  1. Reducing asthma burden.
  2. Finding the healthiest route for specific destination.
  3. Evaluating the health effect of the Luas Cross City interchange line.

The project aims to play an important role in expanding our understanding of social and environmental determinants of health and informing decision-making to promote public health. It will highlight the potentials of using sensor networks and smart technology to engage citizens, collect data, disseminate findings, and put those findings into meaningful next steps to support public health. This 3-year fellowship will be based at University College Dublin and the CONNECT centre strengthened through industry collaborations with IBM, Intel and Google. In addition to the main research objectives, the fellowship will be aligned to an individual career development plan for the applicant.

  • Project name:  Big Data Driven Geo-Spatial Models Supporting Public Health
  • Researcher: Harutyu Shahumyan
  • Supervisor: Francesco Pilla
  • Institution: UCD (Connect)
  • Duration: 36 Months

Advanced RF/mm-Wave Power Amplifier Research for 5G Cellular Communications

The widespread demand for higher data rates in cellular communication systems has led to introduction of the fifth generation (5G) of these systems. One of the key requirements for 5G is to provide ubiquitous connections with multi-Gbps data transmission to meet 1,000 times increased capacity. This goal is expected to be achieved through using two frequency streams, frequencies lower than 6 GHz (5G sub-6 GHz) where currently are employed by the existing mobile communication systems as well as frequencies above 24 GHz (5G mm-wave) where higher spectrum is available.

There are several challenges in hardware design of 5G cellular systems. Power amplifier (PA) is a key component in transmitter of the communication systems. PA development for 5G entails dealing with several challenges including maintaining linearity and efficiency while amplifying signals with wide modulation bandwidth and high peak to average ratio. Moreover, PA design in mm-wave frequencies is quite challenging mainly due to degraded performance of active and passive devices. 

In this research we are planning to develop new architectures and innovative circuit design techniques for 5G PAs. The research is targeted to cover both sub-6 GHz and mm-wave frequency ranges considered for 5G. The architectures and techniques are specifically developed for integrated circuit (IC) implementation of PAs. Currently, we have developed some preliminary PA architectures based on distributed amplifier that are presented in the proposal. We will design and fabricate several proof-of-concept monolithic microwave integrated circuit (MMIC) PAs using GaN and GaAs MMIC technologies. Also, we have proposed a PA architecture based on combination of three concepts, GaN MMIC output stage, CMOS RF-DAC (digital to analog converter) as digital driver stage, and DPD (digital pre-distortion) system for PA linearization. This part of research will be conducted based on collaboration of two research groups at University College Dublin.

  • Project name: Advanced RF/mm-Wave Power Amplifier Research for 5G Cellular Communications
  • Researcher: Gholamreza Nikandish Modaber
  • Supervisor: Anding Zhu
  • Institution: UCD (Connect)
  • Duration: 36 Months