Sensors have for a long time played a vital role in battle awareness for all our armed forces, ranging from advanced imaging technologies, such as radar and sonar to acoustic and the electronic surveillance. Sensors are the "eyes and ears" of the military providing tactical information and assisting in the identification and assessment of threats. Integral in achieving these goals is signal processing. Indeed, through modern signal processing we have seen the basic radar transformed into a highly sophisticated sensing system with waveform agility and adaptive beam patterns, capable of high resolution imaging, and the detection and discrimination of multiple moving targets. Today, the modern defence world aspires to a network of interconnected sensors providing persistent and wide area surveillance of scenes of interest. This requires the collection, dissemination and fusion of data from a range of sensors of widely varying complexity and scale - from satellite imaging to mobile phones. In order to achieve such interconnected sensing, and to avoid the dangers of data overload, it is necessary to re-examine the full signal processing chain from sensor to final decision. The need to reconcile the use of more computationally demanding algorithms and the potential massive increase in data with fundamental resource limitations, both in terms of computation and bandwidth, provides new mathematical and computational challenges. This has led in recent years to the exploration of a number of new techniques, such as, compressed sensing, adaptive sensor management and distributed processing techniques to minimize the amount of data that is acquired or transmitted through the sensor network while maximizing its relevance. While there have been a number of targeted research programs to explore these new ideas, such as the USs "Integrated Sensing and Processing" program and their "Analog to Information" program, this field is still generally in its infancy. This project will study the processing of multi-sensor systems in a coherent programme of work, from efficient sampling, through distributed data processing and fusion, to efficient implementations. Underpinning all this work, we will investigate the significant issues with implementing complex algorithms on small, lighter and lower power computing platforms. Exemplar challenges will be used throughout the project covering all major sensing domains - Radar/radio frequency, Sonar/acoustics, and electro-optics/infrared - to demonstrate the performance of the innovations we develop.

Optical imaging is perhaps the single most important sensor modality in use today. Its use is widespread in consumer, medical, commercial and defence technologies. The most striking development of the last 20 years has been the emergence of digital imaging using complementary metal oxide semiconductor (CMOS) technology. Because CMOS is scalable, camera technology has benefited from Moore's law reduction in transistor size so that it is now possible to buy cameras with more than 10 MegaPixels for £50. The same benefits are beginning to emerge in other imaging markets - most notably in infrared imaging where 64x64 pixel thermal cameras can be bought for under £1000. Far infrared (FIR), or terahertz, imaging is now emerging as a vital modality with application to biomedical and security imaging, but early imaging arrays are still only few pixel research ideas and prototypes that we are currently investigating. There has been no attempt to integrate the three different wavelength sensors coaxially on to the same chip. Sensor fusion is already widespread whereby image data from traditional visible and mid infrared (MIR) sensors is overlaid to provide a more revealing and data rich visualisation. Image fusion permits discrepancies to be identified and comparative processing to be performed. Our aim is to create a "superspectral" imaging chip. By superspectral we mean detection in widely different bands, as opposed to the discrimination of many wavelengths inside a band - e.g. red, green and blue in the visible band. We will use "More than Moore" microelectronic technology as a platform. By doing so, we will leverage widely available low-cost CMOS to build new and economically significant technologies that can be developed and exploited in the UK. There are considerable challenges to be overcome to make such technology possible. We will hybridise two semiconductor systems to integrate efficient photodiode sensors for visible and MIR detection. We will integrate bolometric sensing for FIR imaging. We will use design and packaging technologies for thermal isolation and to optimise the performance of each sensor type. We will use hybridised metamaterial and surface plasmon resonance technologies to optimise wavelength discrimination allowing vertical stacking of physically large (i.e. FIR) sensors with visible and MIR sensors. We ultimate want to demonstrate the world's first ever super-spectral camera.

Overview: We propose a Centre for Doctoral Training in Data Science. Data science is an emerging discipline that combines machine learning, databases, and other research areas in order to generate new knowledge from complex data. Interest in data science is exploding in industry and the public sector, both in the UK and internationally. Students from the Centre will be well prepared to work on tough problems involving large-scale unstructured and semistructured data, which are increasingly arising across a wide variety of application areas. Skills need: There is a significant industrial need for students who are well trained in data science. Skilled data scientists are in high demand. A report by McKinsey Global Institute cites a shortage of up to 190,000 qualified data scientists in the US; the situation in the UK is likely to be similar. A 2012 report in the Harvard Business Review concludes: "Indeed the shortage of data scientists is becoming a serious constraint in some sectors." A report on the Nature web site cited an astonishing 15,000% increase in job postings for data scientists in a single year, from 2011 to 2012. Many of our industrial partners (see letters of support) have expressed a pressing need to hire in data science. Training approach: We will train students using a rigorous and innovative four-year programme that is designed not only to train students in performing cutting-edge research but also to foster interdisciplinary interactions between students and to build students' practical expertise by interacting with a wide consortium of partners. The first year of the programme combines taught coursework and a sequence of small research projects. Taught coursework will include courses in machine learning, databases, and other research areas. Years 2-4 of the programme will consist primarily of an intensive PhD-level research project. The programme will provide students with breadth throughout the interdisciplinary scope of data science, depth in a specialist area, training in leadership and communication skills, and appreciation for practical issues in applied data science. All students will receive individual supervision from at least two members of Centre staff. The training programme will be especially characterized by opportunities for combining theory and practice, and for student-led and peer-to-peer learning.

AlGaN/GaN high electron mobility transistors (HEMT) are a key enabling technology for future high efficiency military and civilian microwave systems. The aim of this proposal is to provide transformative insight into the underlying physical processes that cause degradation in GaN RF power amplifiers (PA). This is of strategic importance for the UK given its strong RF electronics base, due to the fact that GaN RF power electronics delivers a disruptive step change in systems capability through power densities as high as 40W/mm and frequencies exceeding 300GHz. The UK has internationally leading academic research groups in this field, including Bristol and Cardiff. The key issue addressed in this proposal is that device degradation under RF stress is distinctly different than under DC stress, often resulting in a large increase in source resistance, something that never occurs under DC stress and is not explicable by conventional models. This observation implies that a device in RF operation applies voltage/current stresses, which are inaccessible under static conditions, making it imperative to understand the interaction between the RF operating mode and the degradation mechanism. Bristol has provided seminal contributions to the international effort to understand DC GaN transistor degradation, where an understanding is slowly emerging that includes oxygen related reactions and diffusion processes, and dislocation linked breakdown in GaN transistors. This includes electroluminescence imaging for detection of leakage pathways, dynamic transconductance and transient analysis to detect trapping states, and the simulation of the effect of pulsed operation on bulk and surface traps. Over the last 15 years, Cardiff has established a world leading capability in RF PA design and measurement. In particular waveform engineering systems enable RF current/voltage waveforms to not only be measured directly but also to be manipulated almost at will. This manipulation of the waveform has allowed Cardiff to make seminal contributions to the understanding of high efficiency RF PA operation. In this project, the unique capability to 'tune' RF operation into extremely well defined states to enable 'controlled' RF stressing will be used to gain the step change understanding of RF device degradation. Reverse engineering of failed devices, electrical and electro-optical measurement before/after and during the RF stress, combined with physical device simulation, will be used to determine the RF specific degradation mechanisms. This capability to predict, engineer and measure the RF waveforms is key to achieving an understanding of the RF stresses that devices undergo during PA operation, and then to determine and specify the safe-operating-area for HEMTs. This project utilises a partnership with state-of-the-art foundries in Germany and the USA, allowing the project to use production quality devices, essential for the relevance of the work. The project will be guided in terms of its relevance through guidance and interaction with Selex for systems level issues and IQE for the materials. The key synergy of Bristol and Cardiff will address a vitally important issue for the uptake of this disruptive technology, the identification of the RF degradation mechanisms. This will enable the impact of different modes of RF operation to be predicted, and a novel robust RF reliability test methodology to be developed, thus delivering large UK benefit and international impact.

Graphene is a single layer of graphite just one atom thick. As a material it is completely new - not only the thinnest ever but also the strongest. It is almost completely transparent, yet as a conductor of electricity it performs as well or even better than copper. Since the 2010 Nobel Prize for Physics was awarded to UK researchers in this field, fundamental graphene research has attracted much investment by industry and governments around the world, and has created unprecedented excitement. There have been numerous proof-of concept demonstrations for a wide range of applications for graphene. Many applications require high quality material, however, most high quality graphene to date is made by exfoliation with scotch tape from graphite flakes. This is not a manufacturable route as graphene produced this way is prohibitively expensive, equivalent to £10bn per 12" wafer. For high quality graphene to become commercially viable, its price needs to be reduced to £30-100 per wafer, a factor of 100 million. Hence graphene production and process technology is the key bottleneck to be overcome in order to unlock its huge application potential. Overcoming this bottleneck lies at the heart of this proposal. Our proposal aims to develop the potential of graphene into a robust and disruptive technology. We will use a growth method called chemical vapour deposition (CVD) as the key enabler, and address the key questions of industrial materials development. CVD was the growth method that opened up diamond, carbon nanotubes and GaN to industrial scale production. Here it will be developed for graphene as CVD has the potential to give graphene over large areas at low cost and at a quality that equals that of the best exfoliated flakes. CVD is also a quite versatile process that enables novel strategies to integrate graphene with other materials into device architectures. In collaboration with leading industrial partners Aixtron UK, Philips, Intel, Thales and Selex Galileo, we will develop novel integration routes for a diverse set of near-term as well as future applications, for which graphene can outperform current materials and allows the use of previously impossible device form factors and functionality. We will integrate graphene for instance as a transparent conductor into organic light emitting diodes that offer new, efficient and environmentally friendly solutions for general lighting, including a flexible form factor that could revolutionize traditional lighting designs. We will also integrate graphene into liquid crystal devices that offer ultra high resolution and novel optical storage systems. Unlike currently used materials, graphene is also transparent in the infrared range, which is of great interest for many sensing applications in avionics, military imaging and fire safety which we will explore. Furthermore, we propose to develop a carbon based interconnect technology to overcome the limitations Cu poses for next generation microelectronics. This is a key milestone in the semiconductor industry roadmap. As a potential disruptive future technology, we propose to integrate graphene into so called lab-on-a-chip devices tailored to rapid single-molecule biosensing. These are predicted to revolutionize clinical analysis in particular regarding DNA and protein structure determination.