Search

Visual sensors are key to robot perception, which can not only help robot localisation but also enable robots to interact with the environment. However, in new environments, robots can fail to distinguish the static and dynamic components in the visual input. Consequently, robots are unable to track objects or localise themselves. Methods often require precise robot proprioception to compensate for camera movement and separate the static background from the visual input. However, robot proprioception, such as \ac{IMU} or wheel odometry, usually faces the problem of drift accumulation. The state-of-the-art methods demonstrate promising performance but either (1) require semantic segmentation, which is inaccessible in unknown environments, or (2) treat dynamic components as outliers -- which is unfeasible when dynamic objects occupy a large proportion of the visual input. This research work systematically unifies camera and multi-object tracking problems in indoor environments by proposing a multi-motion tracking system; and enables robots to differentiate the static and dynamic components in the visual input with the understanding of their own movements and actions. Detailed evaluation of both simulation environments and robotic platforms suggests that the proposed method outperforms the state-of-the-art dynamic SLAM methods when the majority of the camera view is occluded by multiple unmodeled objects over a long period of time.

Hepatocellular carcinoma (HCC) constitutes a major disease burden worldwide. Much of its high mortality-to-incidence ratio can be attributed to late diagnosis, resulting in poor survival. Motivated by the need to develop a novel non-invasive diagnostic test to improve the chances of early diagnosis, this thesis makes use of metabonomic technologies to discover and validate metabolites as potential novel diagnostic markers for HCC. First, a systematic review of the literature on the topic was conducted to collate all published evidence of metabolites that were reported to be discriminatory for HCC. A bespoke risk of bias assessment tool was developed for metabonomic studies and a weighted score system was implemented to rank metabolites based on the strength of evidence. This resulted in a ranked list of metabolites with the greatest potential to be followed up for validation for each of the sample types (tissue, blood and urine). Then, validation of urinary metabolites with the greatest potential concluded from the systematic review was performed using data acquired from a UK cohort. None of the previously reported difference between HCC and cirrhosis groups could be reproduced, indicating the current lack of candidate markers specific for HCC detectable in urine. Finally, an exploratory analysis of serum 1H-NMR data from a UK and a Nigerian cohort was performed. Common and different alterations in metabolite levels between the two cohorts were compared. Glutamine-to-glutamate ratio was identified as a potential marker with the best discriminatory power between HCC and cirrhosis patients and this was validated using an independent cohort from the Gambia. This work adds to the ongoing effort to elucidate metabolites with the best potential to be further validated with the goal of developing a novel diagnostic test for HCC.

Tracing the fate of fast irreversible chemical reactions remains a challenging and intriguing task in chemistry. While slow reactions can be monitored by routine NMR techniques at ease, real-time monitoring of fast reactions that are initiated by mixing, however, poses challenges to traditional NMR experiments and therefore requires specialised techniques. A state-of-theart stopped-flow piece of equipment was utilised in conjunction with a high-resolution NMR spectrometer, to enable the data collection of rapid reactions including identification of intermediates, side-products, and catalysts, and quantification of selectivity. This work consists of three main aspects regarding the stopped-flow technique: i) evaluating the physical parameters and functionalities of the stopped-flow instrument; ii) monitoring rapid irreversible reactions with half-lives shorter than seconds; and iii) exploiting the non-pre-magnetised input channel of the stopped-flow instrument for magnetisation transfer studies. Thus far, effective kinetic data collection for irreversible chemical reactions of half-lives in the order of milliseconds was achieved, with the aid of interleaved stopped-flow experiments. This was demonstrated by the base-catalysed protodeboronation of pentafluorophenyl boronic acid, of t1/2 = 45 milliseconds at 313 K. The functionality of the non-pre-magnetised stopped-flow input channel was experimentally validated by the chemical exchange studies without the necessity of isotopic labelling. In addition to the stopped-flow technique, a closely-related yet independent work on the measurements of longitudinal relaxation time constant (T1) was conducted. T1 is an important NMR parameter essential for quantitative NMR experiments and molecular dynamic studies. Therefore, prior to kinetic analysis, it is necessary to measure the T1 of spins in the system. Comparing to the most conventional approaches such as the 'Inversion Recovery' experiments, the methodologies developed in this work showed significant improvements in time efficiency and achieved sufficiently accurate T1 measurements. Furthermore, it was proved that continuous heteronuclear decoupling could be applied in T1 measurements, affording sensitivity enhancements from the removal of Jcouplings and the Nuclear Overhauser Effects. This application is particularly beneficial for insensitive nuclei, which intrinsically require longer experimental time. Throughout the developments of T1 methods with heteronuclear decoupling, a range of decoupling schemes were evaluated to highlight the appropriate choice for different spin systems.

Dependency parsing has played a pivotal role in NLP over the last few decades, and non- projective, graph-based dependency parsers have become a dominant approach for this task in the 21st century. While much research has focused on advancing log-linear parameterisations and neural network development, there has been a paucity of research addressing the fundamental algorithms that allow us to use graph-based dependency parsers. This thesis examines algorithms used in four stages of non-projective, graph-based dependency parsers: inference, sampling, decoding, and significance testing. The thesis will guide the reader through a series of novel extensions to existing algorithms, as well as the original develop- ment of new algorithms concerning these four processes in the dependency parsing galaxy. We provide a framework for efficiently evaluating a family of expectations that can be used for inference, such as risk, the Shannon entropy, the KL divergence, inter alia, by capitalising on the connection between gradients and expectations. We further introduce two fast methods for sampling trees from graph theory, and extend them to satisfy a common constraint in dependency parsing schemes that only allows one edge to emanate from the root. Additionally, we contribute the first sampling-without-replacement algorithm for graph-based dependency parsers. We also analyse, simplify, and extend algorithms for decoding the one-best and K-best trees and discuss existing algorithms and new modifications to enable root constrained decoding. Finally, we propose a novel algorithm to perform an exact paired permutation significance tests when comparing the performance of two parsers. Not only is this the first exact algorithm to perform this test, but it is also faster than current approximation strategies. This thesis offers insights into efficient algorithms in different aspects of dependency parsing. We shed light on core algorithms from the graph theory literature and develop these algorithms so that they execute more efficiently and satisfy necessary constraints in dependency parsing. This hitchhiker's guide to dependency parsing is accompanied by publicly available code that can be easily integrated into modern parsers, enabling more efficient parsers in future research.

This thesis contains three chapters studying asset prices from different financial markets to understand the economic forces driving their movements and recover economic variables of interest. In Chapter 1, I develop and estimate a model to quantify the effects of financial constraints, arbitrage capital, and hedging demands on asset prices and their deviations from frictionless benchmarks. Using foreign exchange derivatives data, I find that financial constraints and hedging demands contribute to 46 and 35 percent variation in the deviations from covered interest parity of the one-year maturity. While arbitrage capital fluctuation explains the remaining 19 percent variation on average, it periodically stabilizes prices when the other two forces exert disproportionately large impacts. The model features general financial constraints and produces a nonparametric arbitrage profit function. I unveil the shapes and dynamics of financial constraints from estimates of this function. In Chapter 2 (co-authored with Ian Martin), we propose a framework to compute sharp bounds of the crash probability of an individual stock using option prices. Empirical tests suggest that these bounds are close to the exact forward-looking crash probabilities. Out of sample, either the lower or upper bound outperforms combinations of stock characteristics in terms of forecasting stock-specific crash events. Applying the framework to study the equity of global systemically important banks (G-SIBs) gives rise to forward-looking fragility and stability measures of the global financial system. In Chapter 3 (co-authored with Jiantao Huang), we develop a transparent Bayesian approach to quantify uncertainty in linear stochastic discount factor (SDF) models. We show that, for a Bayesian decision maker, posterior model probabilities increase with maximum in-sample Sharpe ratios and decrease with model dimensions. Entropy of posterior probabilities represents model uncertainty. We apply our approach to quantify the time series of model uncertainty in North American, European, and Asian Pacific equity markets. Model uncertainty is countercyclical in these markets before the 2008 financial crisis, but remains high afterwards. It predicts investors' asset allocation decisions across equity and fixed-income funds. In survey data, investors tend to be more pessimistic about equity performance during periods of high model uncertainty.

Metal matrix composite (MMC) materials can provide very desirable material and component properties. MMCs can be divided into two groups; those that are reinforced by long continuous fibres (such as carbon, alumina, silicon, boron and carbide) or those reinforced by discontinuous fibres (particulates). The addition of fibres and particulates reinforcements enhance desirable properties of the base metal (matrix). A key benefit of MMCs is the ability to tailor component behaviour via fibre orientation or through variation of fibre dispersion through the component. The strength properties of MMC's are known to be heavily influenced by the manufacturing process, caused by changes in the matrix microstructure and through the introduction of residual stresses. Residual stresses are introduced into the composites due to the thermal expansion mismatch between the fibre reinforcements and the matrix. It is very difficult to determine the internal residual stresses within MMCs. This thesis presents an indirect method to predict the thermal residual stresses, through matching numerical and experimental curvature. This research focuses on predicting manufacturing residual stresses to improve the performance of silicon carbide fibre reinforced aluminium 6061 matrix composites. The initial numerical models presented are for composite plates with [90/0] and [+55/-55] fibre. The materials used for the finite element models are elastic-plastic and temperature-dependent. To test the numerical models' ability to predict the curvature, composite laminates were manufactured to the same specifications using the same manufacturing process conditions to validate with the numerical models. The curvatures between the models and manufactured plates showed a good correlation, giving confidence in the models' ability to predict the residual stresses for more complex components. The validated methods developed were then applied to a pressure vessel (burst tube) to demonstrate what residual stresses would be developed within a real component due to the manufacturing process.

Space planning is an important aspect of architectural design. It is the process by which available space is differentiated into zones for particular purposes. In existing literature, a wide range of computational methods have been applied to space planning with a view to automating the process. However, such research was limited by tractability and knowledge extraction issues, and a lack of research into how designers use automated design tools in practice. To address the tractability and knowledge extraction issues, a general data-driven direction was adopted for this research. In this thesis, contemporary residential flat layout design in the UK was studied as an example problem in space planning. The problem was addressed in two separate tasks. The first task was estimating the feasibility of flat types to obtain a realistic brief for a given footprint. The second task was to generate a layout based on a brief for a given footprint. A prototype program was built for each of these two tasks. Prototype Program 1 used a Random Forest classifier that learned from a dataset and estimated flat types for a given footprint. Prototype Program 2 applied a novel case-based learning framework. It used the shape and context of footprints as an index to retrieve existing cases as grating layouts and then used an optimisation method to create new layouts for the given footprints. To research how designers interact with automated design tools, the final part of this thesis reports the design and results of experiments conducted with a panel of eight designers. The experiment examined human-computer interactions with the designers under two contrasting models using a Wizard-of-Oz technique. The designers' responses were recorded in interviews and analysed in detail. This research contributes to the understanding of data-driven space planning, particularly to the expanded use of grating representation and footprint shape-context as an indexing method. The study of the interaction between human designers and computer software provided insight into vi important issues such as design agency, explainability and design exploration, when integrating automated tools into design practice.

China has been experiencing a series of socio-economic transitions since the market reform was explicitly accentuated in the 1990s. On the one hand, China has grown to be the second largest economy and is now producing one of the largest groups of university-educated workers in the world. On the other hand, as a newly emergent player, the professional service sector has contributed to assimilating highly-educated graduates and accelerating the trends of financialisaton, internationalisation and knowledge-based expansion of Chinese society and economy. However, how the growing professional service sector intersects with Chinese transitions is underexplored. In particular, regardless of its theoretical and empirical significance, how professional service firms recruit graduates has long been neglected by existing literature. In this sense, this thesis aims to remedy this gap through a qualitative investigation into the campus recruitment of professional service firms in China. In Chapter 1, I begin by describing key background and context of this thesis as well as its theoretical framework and methodology. In light of a theoretical framework of symbolic boundaries and positional conflict, it attempts to provide the insights into a site of elite hiring and explore its theoretical potential. Drawing on 73 semi-structured interviews with recruiters in elite financial and consulting firms, contents of job advertisements and an array of on-site observations to campus fairs and a day of job interviews, this study is structured as four empirical chapters that chronologically analyse the process of graduate recruitment. Chapter 2 surveys what kinds of talents recruiters were looking for before the formal selection started. Chapter 3 and 4 focus on how recruiters screened and selected résumés. Chapter 5 probs into how recruiters assessed interview performances. Overall, this thesis argues that graduate hiring was a culturally 'scripted' project of elite professional service firms in China and shaped by recruiters' cultural belief and wider organisational and institutional circumstances. In this context, recruiters tended to use various information in order to select graduates akin to or preferred by them or resonant to their ideals for the purpose of maintaining the order of professional workplaces.

This thesis contributes to the growing literature on the textual analysis of news and the cross-section of stock returns throughout three main chapters based on the Dow Jones Newswire Archive. First, I consider in an empirical setting how the presence/absence of news coverage affects lottery-type stocks, measured by the maximum daily returns (MAX) in the prior month. I find an augmented negative relationship between MAX stocks without news and expected returns, whereby MAX with news coverage generates return momentum. The differing future return relationships between MAX stocks with and without news appears to be best explained by information uncertainty mitigation upon news arrival. Overall, this finding suggests that news plays a role in resolving information uncertainty in the stock market. Next, by applying a "co-coverage" concept, I propose to identify each firm's news-based peers (NBPs) and thereby construct a time-varying firm-centric grouping aiming to augment existing industry classification schemes. The advantage of NBP-augmented schemes over traditional industry classifications is to better capture the up-to-date relative importance of the economic links between a base firm and its peers. I show that the base firm's share price responds more favourably to the return shocks of its NBPs than of its traditional industry peers that are not NBPs. The response persists for several months, suggesting that the up-to-date relative importance of the economic links is not immediately clear to investors. Additional empirical tests show that the persistent response, referred to as NBP momentum, can partially unify several lead-lag return momentums previously documented and is consistent with the investor attention hypothesis. Taken together, these results suggest that monitoring news co-coverage is a key to understanding the lead-lag return momentums documented in the literature. Finally, I distinguish news coverage as being either slowly or quickly incorporated into contemporaneous stock prices. The return spread between stocks classified according to these two types of news yields a statistically significant profit of 139 basis points per month. This abnormal return cannot be explained by other well-known risk factors and is robust to allow for trading costs. Overall, this research refines the role of news regarding information dissemination in the financial markets.

Research on ethnic minorities' education in China has tended to focus on certain minority groups of the country but paid limited attention to the experiences of Mongolians. This study explores Mongolian students' experiences in accessing and participating in Chinese higher education and how their higher learning experiences have shaped their sense of identity. The theoretical framework is based on Bourdieu's notion of field, habitus, capital and the education and social reproduction theory. Furthermore, two theoretical concepts, identity and intersectionality, are applied and related to Bourdieu's perspectives regarding identity and representation. An embedded mixed methods design is adopted that is principally qualitative and supplemented with quantitative data. Survey data and semi-structured interviews were used, including 30 participants from four higher education institutions. The challenges for Mongolians in accessing higher education and their different culture shocks in learning and socialising in universities are strongly influenced by their pre-university educational pathway, which is decided by the current educational structure. This reveals that Mongolian students need to be understood as a diverse rather than homogeneous group based on intersectional features such as their varying educational, social and economic capital. The tension between their original habitus and the new institutional habitus in higher education have impacted Mongolian young people's social, personal, ethnic and national identity development. Educational inequalities are reproduced not only between ethnic minority and ethnic majority students but also inside the Mongolian group due to their multiple social positioning and power relations in the higher education field. This study contributes to the understanding of Mongolians' higher education experiences in China and emphasises the multicultural diversity and potential inequalities in educational access, cultural adaption, and academic achievement of the Mongolian group.

This thesis examines the production and reception of New Countryside TV dramas in China using the theory of neoliberalisation. A review of the current academic debates on neoliberalism, actually existing neoliberalism and China's neoliberalisation indicates that neoliberalism is not a static, singular, and rigidly imposed global experimentation. Rather, it is an ongoing project that is mediated by inherited social and institutional structures. Using neoliberalism as an analytical device, this thesis focuses on illustrating how China's neoliberal transformation intersects with China's media reform process. The communication event of New Countryside TV dramas was selected as the case to explain how the Chinese mass media reproduce and legitimate the Chinese state's rural neoliberal transformation agenda. By examining data collected from documents, archives and semi-structured interviews, this thesis argues that China's neoliberalisation is a state- led, media-mediated process. The state plays a dual role, creating the macro development plan for the country and guiding the mass media to endorse the policies it releases in order to win broad social support. The state influences the activities of the marketised mass media sector in two ways. Firstly, it has developed a complicated set of administrative strategies, which combine institutional regulation, financial subsidies, tax policies, market monopoly, party endorsement, etc., to encourage the media industry to produce more products that are consistent with the Socialist New Countryside Construction policy. Secondly, the state uses media reform to form new drama production organisations, and thereby changes the work routines that media professionals follow. The two new patterns of drama production make media professionals more willing to insert more New Countryside Construction Policy-related elements into drama narratives. The state thus can make its New Countryside Construction Policy known to more villagers. The fieldwork in two types of Northern villages reveals that while the New Countryside dramas can effectively increase poorer peasants' entrepreneurship in the economic sense, in both villages, they are unable to enhance peasants' identification with the party and the state in the political sense. Therefore, if neoliberalism is a governmentality which aims to cultivate subjects who are entrepreneurial in the economic sense and governable in the political sense, New Countryside TV dramas are only able to achieve half of this goal. In conclusion, by examining the institutional environment, the production process and the reception of New Countryside TV dramas, this thesis contributes to the discussions on how neoliberalism is being articulated with local inherited political and social processes during the current social transformation in China.

Indoor positioning is a key enabler for context-aware computing. Applications such as navigation, tracking, monitoring, spatial programming and location-based information sharing require the accurate and timely inputs of users' indoor positions. Signal fingerprinting, as a popular infrastructure-free solution for indoor positioning, conventionally builds radio maps that can be used to position a user indoors by manual surveying. Surveying is typically very costly in terms of human effort. This thesis is about crowdsourcing fingerprints so that traditional surveying tasks can be distributed to individual users, especially in a passive way so that the users do not have to take extra actions in order to contribute. Much existing research is based on the assumption that signals collected from multiple devices have to be unified before being merged into the radio map. For example, the signal strength reported by a new device has to be calibrated to a reference device. However, by empirical experiments conducted on two popular crowd-built heterogenous datasets, this thesis finds that the heterogeneity effect is negligible compared to noise, meaning that the heterogeneity will not be as much problematic as in crowdsourcing as it is in traditional fingerprinting research. Furthermore, using heterogenous fingerprints directly in map building outperforms using homogeneous data because the larger quantity of data compensates for the heterogeneity effect. Pedestrian Dead Reckoning (\GLS{PDR}) algorithms are powerful tools to recover walking trajectories from the inertial data. However, in crowdsourcing, inertial data is typically noisy and, thus, error-prone when processed by traditional PDR algorithms. Crowdsourcing can afford to filter out trajectories that are of poor quality and only merge the trustworthy trajectories. Existing PDR algorithms do not handle suboptimal cases properly, especially when the quality of trajectories deteriorates. This thesis proposes a gravity-verified PDR that exploits the observation that the estimated gravity is a good indicator of the quality of the recovered trajectory. The gravity-verified PDR segments trajectories into trustworthy pieces and discards those that cannot be further handled. This thesis proposed two approaches to match trustworthy trajectories to the floor plan in order to build the radio maps based on the idea of maximising the likelihood of observing the recorded fingerprints along the trajectories using Gaussian Processes. The first approach was verified in an office environment with a corridor as its major landscape feature, and achieved a mean positioning accuracy of 2.47 metres, comparable to 2.65 metres, the accuracy of the traditional fingerprinting. The second approach was verified in a public museum with actual visitors as the data contributor and a particle filter as alignment bootstrap aid, and achieved a mean positioning accuracy of 1.66 metres, comparable to 1.87 metres, the accuracy of the traditional fingerprinting. Both approaches achieved equivalent positioning performance to traditional manual surveying methods. The crowdsourced map grows gradually and may not cover all areas, which may lead to systematic errors in positioning when the fingerprint to be positioned is outside the covered area. This thesis also introduces fingerprint anomaly detection algorithm that distinguishes localisable fingerprints from noisy and uninformative fingerprints. With the anomaly detection, positioning algorithm can report the reliability of the positioning result to the users, improving the user experience. This thesis demonstrated a working prototype of the passive crowdsourced indoor positioning system that can greatly lower the human intervention needed for radio map building and maintenance.

This thesis contains two studies that examine the interaction between corporate governance and firm performance. In the first study, I examine whether board friendliness reduces crash risk. I measure friendliness by the Political Homophily Index (PHI), which captures the similarity of political orientations of managers and directors. We find that firms’ crash risk decreases in political homophily. The results are robust when we instrument the change in PHI by the change in local political homogeneity. Our results suggest that better alignment in political orientations facilitates information sharing, including information on bad outcomes in a timely manner. The effect is more pronounced when firms have stronger corporate governance mechanisms and directors have a stronger incentive to acquire information. In the second study, I examine how the use of relative performance evaluation (RPE) affects industry competition. Using data from the U.S. airline industry, we estimate a dynamic game of competition with heterogeneous firms in an oligopolistic market with the presence of RPE contracts. As is standard, RPE makes CEO compensation less sensitive to market conditions. Therefore, the CEO’s propensity to operate in a given market is determined by a trade-off that arises between the reduction in compensation based on market conditions and the gain from being compared to competing agents. The estimation results show that the use of RPE decreases a firm’s tendency to be active under bad market conditions by 10.1%. Conversely, the tendency to be active rises in good market conditions by 12.4%. These effects are stronger for firms with lower fixed operating costs.

The E. coli RNA degradosome is a complex multi-enzyme machine which is central to the post-transcriptional regulation of the cell. Some of its functions include maturing and processing sRNA, rRNA, and tRNA, as well as degrading mRNA. The key components of the RNA degradosome include the endoribonuclease RNase E, the DEAD-box RNA helicase RhlB, the glycolytic enzyme enolase, and the phosphorolytic exoribonuclease PNPase. The degradosome has also been previously shown to associate with the RNA chaperone Hfq to form a small RNA guided machinery that targets defined transcripts. This thesis attempts to investigate several characteristics of the degradosome, including the importance of RhlB, the structure of a portion of the RNase E C-terminal domain that recruits enolase and helicase, and the association between Hfq, ChiX and the RNase E C-terminus. The thesis also explored structural details regarding the small domain of RNase E involved in both RNA binding and oligomerisation and its relationship to the RNA binding KH domains. Utilizing a point mutation in the DEAD box of RhlB, I have found increased RNA affinity to mutant RhlB, the potential structure of Hfq — an RNA chaperone — bound to the sRNA ChiX, and a structural correlation between the RNase E small domain and other KH domains. Preliminary models of Hfq and ChiX structure show a novel binding mode for class II sRNAs as the majority of ChiX associating with the distal face of Hfq. Bioinformatic studies reveal evolutionary roots between the KH domains and the RNase E small domain, supporting the hypothesis that the RNase E small domain may be involved in a novel mode of RNA binding and recognition.

My doctoral research traces the intellectual development of Luisa de Carvajal y Mendoza (1566-1614) by exploring her life and works within the framework of the concept of intellectuality. Rejecting the conventional paths for early modern women of either marriage or convent life, she decided to devote herself to the cause of the faith in an unconventional way and travelled from Catholic Spain to Protestant England where she undertook missionary work and founded her own independent female community. Compared to existing scholarship which has shown a marked tendency towards biographism, my study represents a particular biographist reading by examining Carvajal from an unprecedented perspective based on the concept of intellectuality. This analysis will, on the one hand, explore the concept of intellectuality with the example of Carvajal and, on the other, shed new light on her life and oeuvre with the theory of intellectuality. Based on the idea that knowledge, autonomisation, isolation, and political involvement represent four consecutive steps in the development process of intellectuality, my thesis argues that each of the genres through which Carvajal expressed herself—autobiographical, poetic, and epistolary—not only marks various chapters of her life, but also and more importantly the different stages of her intellectual evolution. In her autobiographical writings, Carvajal demonstrates her erudition and literary skills by representing an identity that is marked by plurality and diversity. She therefore set foot on the path to her autonomisation. Carvajal’s poetry, written in isolation in the attic of her uncle’s house, reflects an epistemological process from individual religiosity towards religious individuality which culminates in her intellectual revelation. This cognitive awakening was subsequently applied to society and manifested itself in her English mission. Far from pursuing a missionary life and the death of a martyr, Carvajal in fact strived to realise her goal of founding and leading a beaterio by provoking Spain’s protection through strategic acts of religious radicalisation, while rebelling against her submission to any authority. Her lifelong goal to found a female community could be understood as the embodiment of her desire for freedom and intellectuality.

Atherosclerosis is a progressive vascular disease, in which the arterial wall thickens compared to its healthy condition. Percutaneous coronary intervention is a minimally invasive, endovascular procedure to treat atherosclerosis, for which a stent is usually deployed at the diseased part of artery to open the blocked vessel. However, the stenting procedure may cause injury (damage) to the vessel wall, resulting in neointimal hyperplasia which is the major contribution to in-stent restenosis. This thesis aims to assess the stenting-induced arterial damage and its contribution to tissue growth after percutaneous coronary intervention. In particular, a range of model subroutines were developed to study the effects of stent-, procedure- and vessel-related factors on tissue damage in the media layer and the development of in-stent restenosis. Using the finite element method, this study presents a mechanistic approach to evaluate the tissue damage caused by stenting and the development of in-stent restenosis in an artery following stent implantation. Hyperelastic models with damage, verified with experimental results, are used to describe the level of tissue damage in arterial layers and plaque caused by stenting. A tissue-growth model, associated with vessel damage, is adopted to describe the growth behaviour of a media layer after stent implantation. Change of lumen diameter with time is used to quantify the development of in-stent restenosis in the vessel after stenting. The effects of pre- and post-dilations on the stenting outcomes are investigated for a polymeric stent. Patient-specific modelling of stent deployment is also simulated to study the effect of stent overlapping on the outcome of percutaneous coronary intervention. The simulation results showed that the larger the artery expansion achieved during inflation the higher the damage introduced to the media layer, leading to an increased level of in-stent restenosis. Softening of the plaque and the artery, due to the pre-dilation, can facilitate the subsequent stent-deployment process, but increases the rate of in-stent restenosis. The post-stenting dilation is effective to achieve further expansion of the artery but causes additional damage to the artery and an increased rate of in-stent restenosis. The plaque in the patient-specific artery is proved to be heavily calcified according to simulation results and the optical coherency tomography image. Single stent produced smaller lumen area, less damage to the media layer and a lower rate of in-stent restenosis than overlapped stents. While non-overlapped stents produced smaller lumen area when compared to overlapped stents, but caused more damage to the medial layer and a higher rate of in-stent restenosis. In summary, this research demonstrated that stent designs and materials strongly affect the stenting-induced damage in the media layer and the subsequent development of in-stent restenosis. The linear correlation of the stenting-induced damage in the media layer and the rate of in-stent restenosis can be used to predict the level of in-stent restenosis developed after percutaneous coronary intervention, including complex clinical procedures such as stent overlapping. Both the pre- and post-dilations increase the rate of the in-stent restenosis. To minimise the in-stent restenosis, a non-compliant balloon is recommended for post-dilation when required. Additionally, to balance the postprocedural and long-term outcomes of stenting, the simulation results recommend the use of a single stent or multiple stents with minimum overlapping for treatment of a long lesion.

Traditionally, to stabilize an unstable continuous-time stochastic differential equation (SDE) by feedback control, continuous-time observations of the system state are required. This is obviously expensive and unrealistic, so recently Mao discretized the observations. This thesis is to investigate the stabilization problem of continuous-time differential equation systems by deterministic and stochastic feedback controls based on discrete-time observations. This problem includes determining the conditions for original system and controller, and calculating the upper bound of the observation interval, namely the minimum of the observation frequency. The SDEs discussed in this thesis are all in the Itô sense. The main mathematical fundamentals used are Itôs formula, Lyapunov's second method and inequalities. The problem was investigated under Lipschitz continuity and linear growth condition. Firstly, I investigated the hybrid SDEs, which is also known as stochastic differential equations with Markovian switching. Using discrete-time observations of system state and mode, we can achieve pth moment stabilization in the sense of asymptotic and exponential stability for p > 1. Our new theory expands from the second moment to pth moment and reduces the observation frequency. Secondly, I used stochastic feedback control, which is based on Brownian motion, to stabilize non-autonomous linear scalar ODEs as well as nonlinear multidimensional hybrid SDEs. Almost sure exponential stabilization is discussed. The new established theory expands the scope of applicable original unstable systems from autonomous ODEs to non-autonomous ODEs and hybrid SDEs. Thirdly, by making full use of the time-varying system property, I used the timevarying observation intervals instead of a constant as before. Non-autonomous periodic SDEs and hybrid SDEs are investigated. Many stabilities are discussed, including asymptotic and exponential stabilities in pth moment for p > 1 and almost surely. My new established theory not only reduces the observation frequencies, but also offers flexibility on the setting of observations.

In this thesis, I study whether stock returns are predictable. Specifically, I study whether the focal firm's linked partners or linked peers can forecast its future returns. The literature of return predictability has found that a lot of forecasting variables, which are mainly constructed by the focal firm's own characteristics, can predict its future returns, but the predictive power from explicitly or implicitly linked firms is not fully explored and understood. The study of interfirm return predictability has become an interesting and important research field, since it challenges current asset pricing theory and models. In this thesis, my research questions are whether inter-firm return predictability exists in the ownership network (namely one new "explicit" network) and whether it exists in the similar employee satisfaction network (namely one new "implicit" network). The overall contribution of the thesis is to find new evidence of return predictability in the inter-firm networks. These new inter-firm return predictabilities are not only an interesting practical fact with implications for investing and hedging, but also have essential implications for new asset pricing factors. In Chapter 2 and Chapter 3, I find the subsidiary-parent return predictability and parent-subsidiary return predictability in a global sample and different regional samples. In Chapter 4, I find that the returns of similar employeesatisfaction-linked firm peers have predictive power over focal firm returns. These results have important implications to call for a new asset pricing model that explicitly incorporates value-relevant information from various inter-firm networks.

In recent years, due to the increasing number of existing and new devices and applications, the wireless industry has experienced an explosion of data traffic usage. As a result, new wireless technologies have been developed to address the capacity crunch. Long-Term Evolution-Licensed Assisted Access (LTE-LAA) is developed to provide the tremendous capacity by extending LTE to 5 GHz unlicensed spectrum. Hyper-dense small cells deployment is another promising technique that can provide a ten to one hundred times capacity gain by bringing small cells closer to mobile user equipments [1]. In this thesis, I focus on three problems related to these two techniques. In Chapter 3, I present a novel activation and sleep mechanism for energy efficient small cell heterogeneous networks (HetNets). In the cell-edge area of a macrocell, the coverage area of a sleeping small-cell will be covered by a range of expanded small-cells nearby. In contrast, in areas close to the macrocell, user equipment (UE) associated with a sleeping small cell will be distributed to the macrocell. Furthermore, the enhanced inter-cell interference coordination (eICIC) technique is used to support range-expanded small cells to avoid Quality of Service (QoS) degradation. Under both hexagonal and stochastic geometry based models, it is demonstrated that the proposed sleeping mechanism significantly reduces the energy consumption of the network compared with the conventional methods while guaranteeing the QoS requirements. Small cells are currently connected to limited backhaul to reduce the deployment and operational costs. In Chapter 4, an optimisation scheme is proposed for small cells to utilise the bandwidth of macrocells as wireless backhaul. I provide the numerical analysis of the performance of both the targeted small cell and the whole network. In Chapter 5, the mobility management (MM) of heterogeneous and LTE-LAA networks are investigated. To avoid Ping-Pong handover (PPHO) and reduce handover failure rate in HetNets, a self-optimisation algorithm is developed to change the handover parameters of a base station automagically. Furthermore, the MM of LTE-LAA networks is analysed. A new handover mechanism is proposed for LTE-LAA networks. Compared with the conventional LTE networks, LTE-LAA networks trigger the handover not only by using UE mobility, but also by the availability of the unlicensed band. A comprehensive analysis of the handover triggering event and handover procedure is presented. Simulation results show that by introducing handover triggered by available unlicensed band, the ratio of handover to unlicensed spectrum has a significant improvement. Therefore, a noticeable enhanced throughput of UEs is achievable by LTE-LAA networks.

Current research about legal modernisation in authoritarian regimes has focused on the utilitarian usage of the law, and most literatures illustrates how the law could be developed under specific demands from the ruling regimes. The KMT (Kuomintang) regime, however, showed a paradoxical path of legal development during its ruling in China: The general framework of a modernised legal system had been established, whereas the ruling of the KMT had been restricted and even undermined. Similar to other authoritarian regimes, under the KMT regime, the motivation factors in conducting legal modernisation included foreign influence, the attempts to centralise power and enhance control over the regions, ambition of promoting social reform via law instrument. However, under the specific social environment of the KMT regime, the direction of its legal modernisation was diverted from other authoritarian regimes. Through the analysis of the KMT's law ideas, we can find out that, apart from the specific social conditions, the KMT's legal ideology played an essential role in the process of legal modernisation. Furthermore, in the KMT's case, although the continuous warfare and weak social infrastructure impeded the effectiveness of legal modernisation, the power struggle inside the top-tier politics, pressure from foreign countries, legal achievements and well-trained legal professionals inherited from the previous regime also contributed to the legal modernisation. These factors combined and made the KMT regime an unique case of legal modernisation among the authoritarian regimes. Based on the KMT's case study, the research on legal development under authoritarian regimes can be enhanced. This study could offer a new perspective in understanding the historical changes in modern China.

This thesis studies how workers are partially insured against business cycle shocks in an imperfect labour market. Business cycle shocks have large effects on the stability of worker’s wage and employment. This study shows these shocks are partially insured by firms, using durable goods, and by social insurance policies. Chapter “Adaptation Costs, Asymmetric Information, and the Business Cycle Effects on the Postgraduate Wage Premium” studies how cyclical wage shocks are insured by long-term contracts provided by firms. I document a new fact that postgraduates have smaller wage shocks than bachelor graduates over the business cycle. I argue the reason for this phenomenon is the adaptation costs, which reduce the value of worker’s outside options and lead workers and firms to agree on a contract with smoother wages. I provide empirical evidence that postgraduates have higher adaptation costs. So postgraduates are better insured by firms and have smaller cyclical wage shocks. Chapter “Asymmetric Information, Durables, and Earnings Shocks” addresses the question of how workers insure themselves against earnings shocks using durable goods. Asymmetric information about the quality of used cars works like a transaction cost, and it implies that used cars are a poor savings vehicle. Chapter “The Impact of Unemployment Insurance on the Cyclicality of Labour Force Participation” studies how business cycle affects worker’s decision on labour force participation, and it shows the level of unemployment insurance benefit plays an important role in shaping fluctuations in the participation rate.

This thesis examines the cultural past of the nɑχɑ people living in Zheba, a rural mountainous area in southwest China. They perceive and live their cultural past through unreconciled fragments that are deeply embedded in their everyday lives. It is a deep past that continuously gives meanings to local people’s existences, even through times when wellbeing in life is most at stake. I trace the nɑχɑ people’s struggles to reclaim the deep past across time through their multi-layered practices of place making, in the experience of the physical body, the domestic environment, the local landscape and beyond. These practices are crucial means to inhabit a world at the margins of the Chinese nation-state. It is a world in which decades of state sponsored violence followed by market changes in the late twentieth century have caused massive social suffering and cultural dislocation. These struggles thus take place against a decades-long background of huge shifts and ruptures in almost every aspect of everyday life. Whilst reshaping local ways of living, these radical transformations have at the same time generated nuanced ways for local people to live with unprecedented uncertainties and contingencies in their life courses. The struggle to balance the tenstion between change and continuity therefore involves unremitting efforts to transmit the fragmented deep past as imperative dimensions of people’s present and future.

The overall aim of this study is to develop a process for manufacturing extra-large aviation components with low residual stresses to minimise their subsequent distortion during the final machining process. The research concentrates on characterising and modelling the residual stress distribution at different stages of the manufacturing process, i.e. solution heat treatment (SHT), quenching and the subsequent residual stress relief. In addition, an effective method has been developed for manufacturing T-shaped components with low residual stresses. Compression tests were conducted to determine the temperature dependent mechanical properties of AA 7050 aluminium alloy. Based on the experimental results, the visco-plastic material behaviour was characterised for a range of test temperatures and strain rates. A dislocation-based constitutive model, describing the thermo-mechanical response of AA 7050 Al alloy in a condition corresponding to the manufacturing process, was developed. Values of material constants in the model were calibrated from isothermal compression test results. To quantify the residual stress evolution in the material during the manufacturing process, residual stress measurements were performed using neutron diffraction, X-ray diffraction and contour techniques on quenched and quenched & cold rolled specimens. The experimental data was used to verify finite element simulations of the residual stress mitigation processes, enabling residual stress predictions to be made. To study the effect of cold rolling on laboratory scale T-section components, a roll assembly was designed and manufactured. According to the measurement results, after cold rolling, the compressive stresses due to quenching at the surfaces of material were removed, though tensile residual stresses were induced near the surface. Generally, the stress predictions from the FE models were in close agreement with the measurement results which show that the FE results are reliable. Feasibility studies were performed to mitigate quench-induced residual stresses in extra-large T-section component. An integrated model was built to predict the effectiveness of cold rolling on residual stresses in heat-treated samples. The optimal parameters (incl. rolling deformation ratio and roll diameter) to minimise residual stresses in scaled-down T-section panels were predicted. Based on these simulation results, the integrated numerical models of cold compression and rolling on residual stress distribution in large-sized T-section components were developed. To reduce the residual stress in large-sized heat-treated T-section components, this study showed that, although cold compression without overlap can effectively relax residual stress in the material, for the case of multiple compressions with overlap regions, the residual stresses distribution post-compression are still sufficient to lead to component distortion. The cold rolling technique is a more cost-effective cold working technique to improve the residual stress distribution in T-profile large-sized components, compared with multiple cold compressions technique. For the cases considered, for reducing residual stresses in extra-large T-section components the FE models also indicate that the optimised parameters to relax residual stresses in the component is a deformation ratio of 1.5 % via a set of rolls with 400 mm radius. These results provide a valuable guideline for future industrial production activities.

Directional change (DC) is a new concept for summarizing market dynamics. Instead of sampling the financial market at fixed intervals as in the traditional time series analysis, by contrast, DC is data-driven: the price change itself dictates when a price is recorded. DC provides us with a complementary way to extract information from data. The data sampled at irregular time intervals in DC allows us to observe features that may not be recognized under time series. In this thesis we propose our new method for the summarizing of financial markets through the use of the DC framework. Firstly, we define what is the vocabulary needed for a DC market summary. The vocabulary includes DC indicators and metrics. DC indicators are used to build a DC market summary for a single market. DC metrics help us quantitatively measure the differences between two markets under the directional change method. We demonstrate how such metrics could quantitatively measure the differences between different DC market summaries. Then, with real financial market data studied using DC, we aim to demonstrate the practicability of DC market analysis, as a complementary method to that of time series, in the analysis of the financial market.

In the recent years many investigations into combustion instability phenomena have been undertaken. This has been partly driven by the development of lean burn gas turbine combustion systems which are more prone to such instabilities. Most of this previous research was based on the assumption that the effect of liquid fuel flow fluctuations could be ignored. However, work on active combustion control has shown that even small fuel flow variations can be of significant effect. In this thesis, focus is placed on the transient response of the liquid fuel flow issuing from an injector to imposed fluctuations. Flow models that are appropriate to modelling the transient fluid response based on the time scales being considered are firstly introduced and categorized. A fully resolved injector network model was developed according to the actual complex geometry of the injector fuel gallery system. This was validated using rigid column theory and was compared with the experimental results. The elastic and rigid column theory results obtained from this complex model showed that the injector fuel gallery system can be considered as a rigid column. In other words fluid compressibility effects within the injector can be ignored. However, the compressibility of the liquid should be included when the complete fuel injector system (i.e. including a fuel supply pipe) is considered. A further simplification of the fuel injector was conducted based on the conclusion above in which the complex fuel injector geometry was replaced with simple resistive and inertial components. This simplified injector model was evaluated at three operating conditions that corresponded to different operating mass flow rates. The boundary condition effects were investigated using the simplified injector model, with quarter wave and half wave affects being observed in the fuel supply pipe leading to significant unsteady fuel flows through the injector. This boundary condition has no effect on the injector transient coefficient, but does affect the system response. An experimental scheme was developed to measure the transient coefficients of varies injector geometries. Four simplified injector geometries have been tested. Initially a pressure perturbation was imposed on the face of the injector. However, unexpected negative derived inertial lengths were obtained due to mechanical vibration of the test rig. Subsequently improvements were made with upstream pressure perturbations being introduced into the fuel supply system. Using this methodology, acceptable derived inertial injector lengths were obtained over a limited frequency range. The measured inertial lengths of the simplified injector were in reasonable agreement with the injector geometry (allowing for end correction effects) up to a frequency of approximately 300Hz. Similarly, the measured values agreed with the predicted values. In this way, the injector resistive (based on the mean pressure drop) and inertial (based on the injector inertial length) values can be determined.

In this thesis, the development of a new methodology for the C–H oxidation of a variety of organic substrates using a bimetallic vanadium-titanium molecular sieve, VTi AlPO-5, is presented. Preliminary studies using model substrate propylbenzene were conducted to elucidate trends in catalytic activity by varying a wide array of reaction and catalyst parameters. Following that is a series of studies documenting the uses of VTi AlPO-5 in the oxidation of other substrates, in which different types of reactivity were probed. The catalyst is highly active and capable of redox- as well as acid catalysis due to the inclusion of its vanadium and titanium centres, the combination of which has been found to be capable of effecting two-step syntheses in the same pot, or enhancing selectivity to very high levels in several reactions. The physicochemical nature of the material as well as spectroscopic investigation of the V and Ti centres in the catalyst has been investigated using a variety of techniques. This data, combined with the catalytic data obtained during reactions, led to the proposal of structures for the metal ions and a preliminary reaction mechanism.

Background: Neurofibrillary tangles (NFTs) is one of the two ‘hallmarks’ of Alzheimer’s disease (AD). The main constitute of these tangles are known as the paired helical filaments (PHFs), which are essentially abnormal aggregate of microtubule associated protein Tau. It has been known for many years that the hyper-phosphorylation of protein Tau is a persistent feature in NFTs. However, whether hyper-phosphorylation of Tau is a cause or an effect to aggregation remains elusive. Mass spectrometry can now serve as an alternative to immunoblotting analysis for post translational modifications (PTM) study. Results: We developed a mass spectrometry method targeting on Tau protein for phosphorylation study. For a decent MS characterisation run, HEK 293 cells confluent in 5-10 plates (50 cm2) are required to be fed to the sample processing procedure. Typically, it takes 4-6 days to prepare the cells into MS samples. The targeting approach was able to discover peptides less intense by two orders of magnitude than that spotted using the general discovery method. The application of the targeted method on the soluble Tau recovered from HEK 293 cells allowed the solid discovery of three additional Tau phosphopeptide (only one phosphopeptide was discovered previously). Furthermore, for the development of the MS targeted method and high purity MS sample preparation, an efficient recombinant Tau protein production (in E.coli) protocol and a purification workflow for phosphorylated and untagged soluble/functional Tau protein (from HEK 293 cell line) was developed respectively. Conclusion: The Tau protein targeted MS method allows solid identification of additional phosphopeptides at two orders of magnitude lower peak intensity. The compatible purification protocol for soluble Tau could be applied to mammalian cells and tissue that expresses wild type and untagged Tau protein.

In the UK, 12,000 patients die from metastatic breast cancer annually. There is therefore an urgent need to identify the molecules that cause metastasis. Recent work has revealed a role for the RUNX family of transcription factors in the development of metastatic breast cancer. The RUNX proteins form active transcription factor complexes when bound by the heterodimeric partner CBFβ to regulate the expression of metastatic genes. Previous work from our laboratory has demonstrated that knockdown of CBFβ resulted in a decreased invasion capacity of the metastatic breast cancer cell line MDA-MB-231. Three-dimensional culture of MDA-MB-231 cells revealed that loss of CBFβ induces a mesenchymal to epithelial transition (MET). The aim of this project was to determine the role of the RUNX/CBFβ complex in maintaining the mesenchymal phenotype of metastatic breast cancer cells. The data presented show that the phenotype changes were accompanied by changes in EMT marker-gene expression, including Snai2, MMP9, and MMP13. Induction of CBFβ in the CBFβ-knockdown cells remarkably restored both the invasive capacity and the mesenchymal phenotype. Further analysis revealed that maintenance of the mesenchymal phenotype was dependent upon both CBFβ-partner proteins, RUNX1 and RUNX2. Taken together the data presented in this thesis demonstrate that RUNX/CBFβ complexes drive the epithelial to mesenchymal transition (EMT) in breast cancer cells. These findings are likely to be important in the development of potential therapies to inhibit the metastatic spread of breast cancer.

Chinese Ceramics have been among the most important archaeological findings in the study of trade in the Indian Ocean from the 8th to the 19th centuries. They have the advantages of commonality, durability, identity and being unearthed in large quantities. Chinese ceramics provide clues for understanding trading trends and linking the Chinese production industries to consumption markets in the Indian Ocean. However, it seems that their crucial importance for field archaeology in the western Indian Ocean has not been well established, due to the lack of a comprehensive overview of Chinese traded ceramic archaeology and a systematic classification. The thesis is concerned with how Chinese trade ceramics impacted on maritime trade in the western Indian Ocean from the 8th to the 16th century. Based on an archaeological report collecting data from 140 ceramic kiln sites in China and on archaeological ceramic material collected from 129 coastal sites and collections in the western Indian Ocean, this thesis has reviewed the archaeological and chronological development of Chinese trade ceramics. A systematic classification of Chinese trade ceramics in the western Indian Ocean has been developed and built, introduced with a review of the long-term history and researched using quantitative methods.

The use of total ankle replacement (TAR) for treatment of arthritis is rapidly increasing, but survival rates are of major concern. The primary indication for TAR revision is implant loosening, which is linked with inadequate primary stability manifested in higher levels of initial implant-bone micromotion. Finite-element (FE) modelling has been utilised to assess micromotion of arthroplasty implants, but not TAR. Additionally, the biomechanical consequences of TAR malpositioning during surgery – previously linked with higher failure rates – remain unexplored. The aim of this thesis was therefore to apply FE modelling to estimate implant-bone micromotion and peri-implant bone strains of current TAR designs under optimally-positioned and malpositioned cases, and thereby identify fixation features and malpositioning scenarios that place the reconstructed ankle at risk of early loosening. Computational models simulating commonly-used TAR designs (BOX®, Mobility® and Salto®) implanted into the tibia and talus were developed; the loads applied were the contact forces acting in the ankle during gait, as calculated using a previously-validated musculoskeletal model, while implementing muscle-architecture data obtained through dissections of cadaveric legs. Micromotion and strain outcomes were larger for the tibial compared with the talar components, in agreement with previous clinical observations. The tibial Mobility® and talar Salto® components demonstrated the largest micromotion. A gap between the tibia/talus and implant component resulted in a considerable increase in implant-bone micromotion and peri-implant bone strains; the Salto® design was relatively ‘forgiving’ for such malpositioning. It was concluded that better primary stability can be achieved through fixation nearer to the joint line, while relying on more than a single fixation peg, and preserving more of the cortical sidewalls of the bone; incomplete seating of the implant on the bone increases the risk for TAR failure. The models presented in this thesis may assist implant designers and surgeons in improving TAR designs and surgical techniques.

To provide time-critical early warnings in a Tsunami Warning System (TWS), the time delay for both the sensor data exchange process (upstream) and the warning message dissemination processes (downstream) should be minimal, maximising the time available for accurately analysing the situation and giving more time for people in the affected region to react to the warnings. Publish/Subscribe Message-oriented Middleware (PSMOM) in combination with a novel use of a federated broker (broker overlay) can be deployed in TWS to support both time-critical and resilient communication. PSMOM can better manage message bursts caused by a sudden increase in sensor data exchange frequency or by additional sensors coming online. PSMOM can better manage the decrease in available system resources (bottlenecks) caused by a disruption in the underlying network infrastructure (limited resource case). Otherwise, these burst and bottlenecks can cause some brokers to become overloaded, which may in turn degrade the overall system performance and delay decision-making. Existing PSMOM load management solutions have two key limitations when applied to TWS. First, existing work does not consider the message delay requirements for the redistribution and offloading phases of load management. Here, some data is only useful or valid for a short time-span (from tens of seconds to tens of minutes); hence, it needs to be exchanged within this maximum allowed end-to-end transmission delay. Time critical subscribers need to be de-prioritised from being offloaded, as the offloading processes take some time to complete, introducing unexpected delays to message exchange. Second, existing solutions assume that there are surplus system resources for offloading, i.e., less loaded brokers can accept loads from overloaded brokers. However, in a TWS, the underlying network infrastructure may be disrupted which in turn reduces the system capacity. It always takes time to recover from the limited resources situation and during that time, brokers may not have enough system resources to accept loads from overloaded brokers, which may result in total failure of the overloaded brokers. A novel load management framework called ePEER is proposed that extends an existing messaging system, Publish/Subscribe Efficient Event Routing (PEER), with the following main contributions. First, for the surplus resource case, the message delay requirements for different subscription services are considered in the load analysis process when offloading the load to different brokers. Second, for the limited resource case, a feedback driven congestion control mechanism can be used when the underlay network infrastructure is damaged, reducing the available bandwidth of PSMOM. This mechanism limits the publication rate of messages with less value, to better maintain the quality of experience (QoE) of subscribers for the more important messages ePEER is validated with emulation-based experiments. The results show that ePEER outperforms the state of the art load management solution used by PEER: through preventing unnecessary delays introduced to time critical services, and through ensuring important messages can be more efficiently exchanged to improve the QoE of subscribers.

Making headway in the field of biomaterials research and specifically in vivo drug delivery using nanoencapsulation is certainly a challenge but the rewards are obvious when considering potential applications in cancer therapy, drug delivery and uses in the potential in the cosmetic industry. As multi-purpose agents nanocapsules not only require a high success rate of encapsulation but also a remarkable stability for targeted drug delivery, viral vector and emulsion nanocapsules which are considered a great match for this purpose. In this research nanocapsules were fabricated using a layer-by-layer technique: polymer layers are alternatively deposited using opposite charges on a bio-core surface. Smart polyelectrolyte nanocapsules are produced that can be successfully uptaken by cells and catalysed to release drugs at a target site. Previous research in this area attempted to use nanocapsules in the range of 600nm resulting in a high rate of cell fatality during uptake. We focus on three projects all in the range of 100nm, a size range predicted to have an even higher success rate for: 1) Oil-dye assisted emulsion nanoencapsulation. 2) Polystyrene latex beads nanoencapsulation. 3) Adenovirus nanoencapsulation. The characterization of nanocapsules included Confocal Laser Scanning Microscopy (CLSM), Dynamic Light Scattering (DLS), Transmission Electron Microscopy (TEM) and Scanning Electron Microscope (SEM). Polymer layering is examined for stability, successful encapsulation and their biocompatibility. Fabrication smart oil-dye assisted nanocapsules containing water insoluble drugs are using a solvent evaporation to shrink the microcapsules to nanocapsules. With regards to the successful encapsulation of latex beads it is shown that direct polymer deposition on the adenovirus surface could significantly increase the speed of assembly and reduces complexity in fabrication. Adenovirus nanocapsules are examined using the above methods as well as a cell study based on gene expression showing that the capsules successfully infect cells with the virus without causing cell fatalities.

Model Driven Engineering (MDE) is a state-of-the-art software engineering approach, which adopts models as first class artefacts. In MDE, modelling tools and task-specific model management languages are used to reason about the system under development and to (automatically) produce software artefacts such as working code and documentation. Existing tools which provide state-of-the-art model management languages exhibit the lack of support for automatic static analysis for error detection (especially when models defined in various modelling technologies are involved within a multi-step MDE development process) and for performance optimisation (especially when very large models are involved in model management operations). This thesis investigates the hypothesis that static analysis of model management programs in the context of MDE can help with the detection of potential runtime errors and can be also used to achieve automated performance optimisation of such programs. To assess the validity of this hypothesis, a static analysis framework for the Epsilon family of model management languages is designed and implemented. The static analysis framework is evaluated in terms of its support for analysis of task-specific model management programs involving models defined in different modelling technologies, and its ability to improve the performance of model management programs operating on large models.

Large optimization problems that involve either a large number of decision variables or many objectives pose great challenges to nature inspired optimization algorithms. On the one hand, nature inspired optimization algorithms suffer from the curse of dimensionality in the decision space. With an exponentially increased volume of the decision space as well as the complexity of the search landscape, it is challenging for nature inspired optimization algorithms to perform efficient search within limited execution time, in particular when the decision variables are non-separable due to the interactions between them. On the other hand, nature inspired optimization algorithms suffer from the curse of dimensionality in the objective space. In high-dimensional objective space, due to the sparse distribution of the candidate solutions, efficient population diversity management and convergence pressure preservation become particularly important. In this thesis, we present several pieces of work to address the above challenges. Firstly, we present two variants of the particle swarm optimization algorithm to tackle single-objective large problems. By enhancing the swarm diversity, the two algorithms are capable of handling problems with as large as 5000 decision variables. Secondly, we present an inverse model based evolutionary algorithm to tackle multi-objective large problems. By building inverse models that map candidate solutions from the objective space to the decision space, the algorithm is able to enhance the computational efficiency in the optimization of bi- or three-objective problems with a large number of decision variables. Thirdly, we present a reference vector guided evolutionary algorithm to tackle optimization problems with many objectives. By decomposing the high-dimensional objective space into subspaces using a set of reference vectors, the algorithm is able to handle problems with as many as 10 objectives. Finally, we present a benchmark test suite which can be used to assess the performance of nature inspired optimization algorithms on problems with both a large number of decision variables and many objectives. The performance of all proposed algorithms have been verified on widely used benchmark problems in comparison with some other state-of-the-art algorithms. Moreover, the proposed reference vector guided evolutionary algorithm has been successfully applied to the optimization of a seven-objective hybrid electric vehicle controller model designed at the Honda Research Institute Europe.

For vehicles whose duty cycle is dominated by start-stop operation, fuel consumption may be significantly improved by better management of the start-stop process. Pneumatic hybrid technology represents one technology pathway to realise this goal. Vehicle kinetic energy is converted to pneumatic energy by compressing air into air tank(s) during the braking. The recovered air is reused to supply an air starter, or supply energy to the air path in order to reduce turbo-lag. This research aims to explore the concept and control of a novel pneumatic hybrid powertrain for a city bus application to identify the potential for improvements in fuel economy and drivability. In order to support the investigation of energy management, system architecture and control methodologies, two kinds of simulation models are created. Backward-facing simulation models have been built using Simulink. Forward-facing models have been developed in the GT-POWER and Simulink co-simulation. After comparison, the fully controllable hybrid braking system is chosen to realize the regenerative braking function. A number of architectures for managing a rapid energy transfer into the powertrain to reduce turbo-lag have been investigated. A city bus energy control strategy has been proposed to realize the Stop-Start Function, Boost Function, and Regenerative Braking Function as well as the normal operations. An optimisation study is conducted to identify the relationships between operating parameters and respectively fuel consumption, performance and energy usage. In conclusion, pneumatic hybrid technology can improve the city bus fuel economy by at least 6% in a typical bus driving cycle, and reduce the engine brake torque response and vehicle acceleration. Based on the findings, it can be learned that the pneumatic hybrid technology offers a clear and low-cost alternative to the electric hybrid technology in improving fuel economy and vehicle drivability.

This thesis presents work on developing advanced controllable continuous-wave (CW) solid-state Raman lasers operating at near-infrared and visible wavelengths for astronomy and medical applications. An adaptive-optics (AO) technique, which is commonly applied in astronomy, was for the first time integrated within self-Raman lasers for improving the Raman laser output performances and wavelengths control. Over the last decade, intracavity solid-state Raman lasers have been widely utilised to extend the spectral coverage of common crystalline laser materials, especially when they are combined with second harmonic generation or sum frequency generation. However, due to the non-elastic nature of stimulated Raman scattering, a significant thermal lensing is generated within the Raman gain medium, which results in parasitic thermo-optical distortions inside the cavity. The magnitude of this effect scales directly with the Raman laser output power and has been identified as the main limitation in power-scaling crystalline Raman lasers. In this thesis, an AO-based feedback control loop system, consisting of an intracavity bimorph deformable mirror, a photodiode sensor and a PC-based control program using a random-search algorithm, has been implemented inside several Raman laser platforms demonstrating its potential for Raman laser output power-scaling. A power improvement of up to ~ 45 % is reported for a Nd:YVO4 self-Raman laser at λ = 1176 nm using an intracavity AO optimisation. Moreover, a frequency-doubled Nd:GdVO4 self-Raman laser in the yellow waveband (λ = 586.5 nm) has also been power-scaled using the intracavity AO technique, achieving a total power enhancement of ~ 41 %. These represent the potential to significantly alleviate the detrimental thermal lens effect and open avenues which will enable solid-state Raman lasers to reach new output power levels. In addition, several CW Nd:YVO4 self-Raman lasers based on the primary and secondary Raman shifts of YVO4 ( 893 cmˉ¹ and 379 cmˉ¹ respectively) were characterised and reported in this thesis. For the first time an on demand wavelengths selection between laser output at λ = 1109 nm and λ = 1176 nm has been achieved with hundreds of milliwatts output power using an intracavity AO control technique. This represents an important step on the way towards automatic wavelength selectable high power Raman lasers.

A novel, non-coaxial soil model is developed in the context of perfect plasticity for the plane strain condition whilst incorporating initial soil strength anisotropy. The anisotropic yield criterion is developed by generalising the conventional isotropic Mohr-Coulomb yield criterion to account for the effects of initial soil strength anisotropy described by the variation of internal friction angles at different principal stress directions. The model is implemented into the commercial finite element(FE)software ABAQUS via the user defined material subroutine(UMAT). The proposed model is used to predict material non-coaxiality in simple shear tests. The non-coincidence of the directions of principal stresses and plastic strain rates can be reproduced. A faster rate of approaching coaxiality is observed when soil yield anisotropy is presented when compared to the model with an isotropic yield criterion. A semi-analytical solution of the bearing capacity for a smooth strip footing resting on an anisotropic, weightless, cohesive-frictional soil is developed based on the slip line method. A good match of the bearing capacity can be obtained between numerical and semi-analytical results. The results show that the vertical load at plastic collapse of a strip footing resting on an anisotropic soil is lower than that on an isotropic soil. The settlement prior to collapse is larger when the non-coaxial assumption is involved; however, no significant impacts can be observed on the ultimate failure load. In addition, the non-coaxial soil model is applied to investigate tunnelling induced displacement. The results are compared with the results from the centrifuge tests performed by Zhou (2015). For equal volume loss, the normalised settlement trough can be improved by adopting the soil anisotropic parameter β as compared to the experimental results. The maximum settlement is larger in light of larger non-coaxial coefficient for the same degree of the stress reduction. The cross-section of the anisotropic yield criterion developed is a rotational ellipse. Other types of the ellipse are possible. In addition, for simplicity we only consider the effect of initial anisotropy without considering induced anisotropy, and only the simple case of perfect plasticity is investigated. It is suggested that in order to capture the soil behaviours under more complex stress paths, the non-linear and anisotropic elasticity should be associated with the current model, and the development of hardening/softening rules is worth investigating.

Neuroendocrine tumours are rich sources of bioactive peptides and potentially-active fragments of proteins. As the cells that constitute such tumours have active regulated secretory pathways, they constantly deliver cocktails of these peptides into the circulation (tumours) or into the culture medium (cells). Many of these peptides may contribute to distinctive clinical tumour syndromes or may be involved in the growth and metastasis of the tumours themselves. This project, the first of its kind, was directed at the molecular inventory of the secretory peptidome of these tumours by application of high-throughput LC/MS/MS techniques. Individual peptides have been structurally-characterized and their parent proteins identified using bioinformatic techniques. They were then sorted into protein types, such as catalytic, structural, regulatory, etc., and comparisons made between individual tumour types to assess numbers of both common and unique components. Selected peptides were chemically-synthesized and subjected to appropriate bioassays. This project provides a body of basic information that may be of use to basic scientists and clinicians.

Simultaneous interpreting requires efficient use of highly domain-specific terminology in the working languages of an interpreter. By necessity, interpreters often work in a wide range of domains and have limited time to prepare for new topics. To ensure the best possible simultaneous interpreting of specialised conferences where a great number of domain-specific terms are used, interpreters need preparation, usually under considerable time pressure. They need to familiarise themselves with concepts, technical terms, and proper names in the interpreters’ working languages. There is little research into the use of modern terminology extraction tools and pipelines for the task of simultaneous interpreting. A few previous studies mentioned the application of corpora as potential electronic tools for interpreters. For instance, Fantinuoli (2006) and Gorjanc (2009) discussed the functions of specific online crawling tools and explored ways to extract specialised terminology from disposable web corpora for interpreters. However, there has not been any empirical study to test how term extraction tools and the use of corpora can help interpreters increase their preparation efficiency and how these technologies and practices influence interpreters’ simultaneous interpreting performance. This study investigates a corpus-based terminology preparation pipeline integrating building small comparable corpora, using automatic term extractors and concordancers. We compared and evaluated several term extraction and concordance tools for Chinese and English, and a single term extractor and a concordancer with comparatively better performance were selected to be used in the empirical study of this research. With training on how to use the tools for interpreting preparation, interpreters are expected to develop the skills to build their own terminology resources and activate relevant terms for specialised simultaneous interpreting tasks. This study also investigates the effect of using the tools on trainee interpreters’ performances by looking at the quality of their simultaneous interpreting outputs. For this purpose, we ran two experiments with MA trainee interpreters at the University of Leeds using different preparation procedures (and tools) to prepare for simultaneous interpreting tasks (English and Chinese, both directions) on two specialised topics: Seabed Minerals (SM) and Fast Breeder Reactors (FR). I also collected data from focus groups to investigate the trainee interpreters’ views on the use of different procedures (and tools). Our results suggest that the preparation procedure using both the term extractor (Syllabs Tools) and the concordancer (Sketch Engine) yielded better preparation results compared with a traditional preparation procedure. It helped improve the trainee interpreters’ terminological performance during simultaneous interpreting by significantly increasing term accuracy scores by 7.5% and reducing the number of omission errors by 9.3%. On the other hand, terminology preparation (through using both the term extractor and the concordancer) is not a “magical cure” for all errors. Our data shows that the preparation procedure (and the tools) only helped to improve the students’ holistic SI scores by 2.8% (but not yielding any statistical significance). This thesis demonstrates that training on terminology preparation for technical meetings could be a useful supplement to the already existing professional interpreting training. It is important for both students and trainers to be aware that electronic tools, when used properly, can assist the interpreters’ terminology preparation and achieve an enhanced performance. It also offers directions for further research in the application of modern term extraction technology for conference interpreters.

With the promotion of demand side responses (DSRs) and low carbon technologies (LCTs), there is a growing interest in visualising the demand information at individual consumer and low voltage (LV) network level, where demands are less aggregated and highly volatile. Yet, traditional load profiling techniques, which are carried out on small data, are struggling to meet the requirements on accuracy and granularity. This thesis contributes to this area by extending traditional load profiling to a big-data context, where refined load profiles (smart data) can be extracted by two novel load profiling techniques for LV networks and individual consumers. The refined load profiles aim to: i) economically visualise LV networks with limited smart-grid monitoring data; ii) transform the smart metering data into a high-detail granular representation of the customers’ daily demand. For the LV networks, this thesis develops a novel concept, LV network templates, which aim to visualise the LV networks in a cost-effective manner. A novel three-stage load profiling method is proposed as: clustering, classification and scaling. By using statistical time-series analysis, three steps are undertaken: i) cluster a vast amount of load data according to their load shapes; ii) classify un-monitored substations to the most similar cluster without sample metering; iii) and also scale them to the right magnitude without sample metering. Through this method, limited representative monitoring data can be used to develop a library of typical load profiles for un-monitored networks, thus saving the cost of extensive monitoring for every single substation. In addition, it is the first load profiling method that can accurately express both load shapes and magnitudes for LV networks. Regarding the customer’s demand representation, the developed time-series analysis needs to be updated due to the volatile and uncertain nature of smart metering data, including inter-related factors such as overall load shapes, sudden spikes and magnitudes. Therefore, an innovative spectral load profiling is proposed to decompose these factors into different spectral levels, characterised by spectral features. By analysing the extracted features on each spectral level separately through multi-resolution analysis, the interference among different factors can effectively be prevented. The proposed method, for the first time, is able to fully capture the energy characteristics at the household level. The developed LV network load templates provide an economical but straightforward way to quantify the available headroom of unmonitored substations over time, providing quantitative information for distribution network operators to integrate LTCs at the minimal costs. The spectral load profiling gives an insight into customer’s energy behaviours with high granularity and accuracy. It can support the customer-specified DSR, tariff design, smart metering validation and load forecasting.

People are the centre of technologies. Understanding, monitoring and tracking the behaviour of people will benefit in various areas including driving assistance, surveillance for safety and caring purposes and applications for machine-people interaction. Particularly, pedestrians attract more attention for two reasons: they restrict the behaviours of people to standing and moving upright; and applications for pedestrian detection and monitoring have positively impact on the quality of life. Pedestrian detection and identification, aims at recognising pedestrians fromstill images and video frames. Together with pedestrian recognition and tracking, this topic attempts to train computers to recognise a pedestrian. The problem is challenging. Though frameworks were designed, various algorithms were proposed in recent years, further efforts are needed to improve the accuracy and reliability of the performance. In this thesis, proposing a modifiable framework for pedestrian identification and improving the performances of current pedestrian detection techniques are particularly focused. Based on appearance based pedestrian identification, a modifiable framework is a novel philosophy of developing frameworks which can be easily improved. For pedestrian identification, a novel protocol where layers of algorithms are hierarchically applied to solve the problem. To compare the detected pedestrians, appearance based features are selected, the "bag-of-features" framework is employed to compare the histogram descriptions of pedestrians. To improve the performances of HOG pedestrian detector, the presence of head-shoulder structure is selected as the evidence of the presence of pedestrian. A novel appearance based framework is developed to detect the head-shoulder structure from the detection results of HOG detector. Furthermore, in order to separate multiple pedestrians detected in one bounding box, a novel algorithm is proposed to detect the approximated symmetry axes of pedestrians.

An information technology that can combine the physical world and virtual world is desired. The Internet of Things (IoT) is a concept system that uses Radio Frequency Identification (RFID), WSN and barcode scanners to sense and to detect physical objects and events. This information is shared with people on the Internet. With the announcement of the Smarter Planet concept by IBM, the problem of how to share this data was raised. However, the original design of WSN aims to provide environment monitoring and control within a small scale local network. It cannot meet the demands of the IoT because there is a lack of multi-connection functionality with other WSNs and upper level applications. As various standards of WSNs provide information for different purposes, a hybrid system that gives a complete answer by combining all of them could be promising for future IoT applications. This thesis is on the subject of `Federated Sensor Network' design and architectural development for the Internet of Things. A Federated Sensor Network (FSN) is a system that integrates WSNs and the Internet. Currently, methods of integrating WSNs and the Internet can follow one of three main directions: a Front-End Proxy solution, a Gateway solution or a TCP/IP Overlay solution. Architectures based on the ideas from all three directions are presented in this thesis; this forms a comprehensive body of research on possible Federated Sensor Network architecture designs. In addition, a fully compatible technology for the sensor network application, namely the Sensor Model Language (SensorML), has been reviewed and embedded into our FSN systems. The IoT as a new concept is also comprehensively described and the major technical issues discussed. Finally, a case study of the IoT in logistic management for emergency response is given. Proposed FSN architectures based on the Gateway solution are demonstrated through hardware implementation and lab tests. A demonstration of the 6LoWPAN enabled federated sensor network based on the TCP/IP Overlay solution presents a good result for the iNET localization and tracking project. All the tests of the designs have verified feasibility and achieve the target of the IoT concept.

Two types of CrAlSiN nano-multilayer coating were designed and synthesised on AISI M2 high speed steel. Selected coatings were also deposited on pre-nitrided M2 substrate to investigate the effect of duplex treatment. A series of type I CrAlSiN coatings, which have sharp interfaces between individual nano-layers, were designed with varying bilayer thickness. An fcc and an hcp phase were identified for all type I coatings, and it has been found that the bilayer thickness has a significant influence on the orientation of both phases, which in turn affects their mechanical properties. Type II CrAlSiN coatings, with blurred interfaces, were synthesised with a medium bilayer thickness. Due to the insignificant difference in composition between neighbouring nano-layers, only one fcc phase was found in type II CrAlSiN coating. The wear resistance of the new nano-multilayer CrAlSiN coating is about 10 times higher than CrN when tested at room and elevated temperatures. This could be attributed to high H/E ratio and good oxidation resistance conferred by the nano-multilayer structure and multi-component composition. It is also found that duplex treatment can improve the wear resistance of nano-multilayer coating by two orders of magnitude due to the improved load bearing capacity of the coating system.

Recently, there has been a dramatic increase in the amount of data transmission within short range local area networks (LAN). Multimode fibre (MMF) is widely used in local area networks because of its coupling and alignment along with the low cost of related components. Graded index MMF has become common due to the reduction in pulse spreading; however, as demands for high bandwidth increase towards a future gigabit rate network, the typical MMF using conventional transmission methods will not be suitable. Meanwhile, this increasing demand for high speed data transmission will soon reach the Shannon capacity limit of single mode fibres. After multiple input and multiple output (MIMO) technology was successfully used in wireless communication, the researcher realised that the same idea could also be applied to an optical fibre network. Optical MIMO techniques are gaining interest in order to create parallel channels over orthogonal modes in a MMF or a few mode fibre (FMF). This approach could lead to a significant increase in the bandwidth distance product and be employed in the next 40Gb/s or even 100Gb/s optical fibre transmission systems. Generally speaking, optical MIMO appears to be the best solution to the bandwidth limitation problem in either short distance MMF or long distance FMF systems. This thesis focuses on designing a simple, cost-effective, and energy efficient optical MIMO system based on MMFs. This proposed system can be realised by combining radial offset launching and annular multi-segment detectors. First, in the initial work, we performed a theoretical and numerical study of the key impairments of MMFs, and the mode propagation in an MMF was analysed mathematically. The variation in electrical field intensity for linearly polarised (LP) modes in the core region of an MMF and the analytical solutions for power coupling coefficients in either radial offset launching or centre launching were presented. In addition, the modal time delays, impulse response, and transfer function were all introduced. Subsequently, the near field intensity pattern (NFP) was simulated at the output facet of the MMF, which indicated that the overall NFP suffered from blurring when it contained mode mixing, and that the intensity pattern was particularly sensitive to the random phase. According to the spatial distribution of the NFP, the annular detector can be exploited more efficiently. All of the results were calculated and plotted using the MATLAB program. Secondly, the optical MIMO model in the multimode fibre was briefly summarised, including the MIMO channel matrix H expression, a mathematical expression of optical MIMO capacity, MIMO channel estimation and an equalization method. Two metrics can be used to characterise the MIMO channel performance: condition number and crosstalk at each receiver. The numerical results demonstrated that the new type of annular multi-segment detector exhibits superior performance compared to the conventional multiple single mode fibre (SMF) detectors, making them attractive for future optical MIMO systems. Finally, the core work of this thesis can be divided into two parts: the modelling of a 10Gb/s intensity modulation direct detection (IM-DD) optical MIMO MMF system; and the modelling of an advanced 10Gb/s coherent differential phase shift keying (DPSK) MIMO FMF system. In both simulation systems, the important transmission parameters of intra-group mode mixing, modal dispersion, chromatic dispersion, and mode attenuation were considered and discussed in detail. In the IM-DD optical MIMO system, the optimization of the transceiver can be based upon the laser spot size and the power flux distribution emitted by the transmitter. Results from the simulation showed that the intra-group mode mixing had a limited impact on system performance, and due to its inability to compensate for linear impairments, the IM-DD optical MIMO was not favourable for long distance transmission systems. Nevertheless, the new type of optical fibre FMF seems to be the most promising candidate for use in long haul transmission systems. Therefore, the well-known DPSK modulation format in conjugation with the coherent detection deployed in FMF was studied. Both heterodyne and intradyne detection schemes were analysed followed by mathematical derivation and numerical simulation; the results illustrated that similar system performances can be achieved in both schemes. Meanwhile, the coherent DPSK simulation results also demonstrated that the linear impairments were almost compensated by the frequency domain MIMO equalization process, which resulted in system performance being independent to transmission distance for up to 10km. This advantage proved that the coherent optical DPSK MIMO system can be employed in long haul networks. As with an IM-DD optical MIMO system, optimization of a coherent MIMO system was also possible. However, in contrast to the optimization of an IM-DD MIMO system, a trade-off had to be made between sufficient spatial diversity at the transceiver and differential modal delay caused by modal dispersion; consequently, the numerical results showed that the proposed coherent optical DPSK MIMO gained reasonable good results without using any active device, such as a spatial light modulator and a mode converter. In conclusion, this proposed optical MIMO system provided easy implementation and integration and is feasible for use in future optical communication systems.

This thesis seeks to analyse military and civilian loss from violence during contemporary armed conflict in order to facilitate understanding of the evolution of war and its impact on human behaviour. It comprises four chapters; the first two concentrate on the 2033 Iraq War whilst the last two are focused upon global armed conflict during the recent past. Chapter 1 explores how and to what extent military deaths during the Iraq war affect US domestic opinion, proxied by various poll questions concerning war-related issues. Having addressed irregular frequencies of poll data that restrict time series application, this chapter renders a fresh perspective on casualty-opinion research, suggesting that cumulative military casualties prior to the poll did not have an immediate effect on the poll respondents' opinion regarding the continuation of military actions in Iraq, Instead, respondents are influenced by marginal casualty information from the previous time period, implying a slow adjustment in forming opinion through the Error Correction Mechanism (ECM). Chapter 2 presents a comparative analysis to gauge any different standards between the US department of Defense and the media in counting violent civilian deaths during the Iraq war. In spite of substantial discrepancies during the initial period of the war, non-parametric tests corroborate that the US military authority and media reports had a non-differential approach towards counting violent civilian deaths during the war period across the spatial and spatiotemporal dimensions. However, the conspicuously conservative count by the US military authority during the initial stage of the war may have hindered the US forces' ability to predict and prepare for the subsequent escalation of violence that brought about large-scale human loss as well as the prolongation of the war which lasted more than 7 years. Chapter 3 analyses to what extent warring actors intentionally used lethal force against civilians, through the employment of a Civilian Targeting Index (CTI), a newly invented measure to indicate the intensity of civilian targeting for each actor. Building upon Chapter 3, Chapter 4 further examines factors that lead to warring actors targeting civilians as opposed to engaging in battle with war combatants. A dynamic panel approach shows that an increase in the degree of civilian targeting in the previous year further intensified civilian targeting in the current year for the actors involved in prolonged armed conflict.