Search

Considering the limitations of latency and power consumption in conventional machine vision systems, this thesis investigates a new visual information process scheme with an emerging visual sensor: the Pixel Processor Array (PPA) by directly processing signals where they are collected, hence avoiding the aforementioned issues with conventional machine vision systems. In particular, this thesis establishes mobile robotic control systems with on-sensor computed results for multiple navigation research. Then, our work investigates novel parallel visual inference approaches, with a particular emphasis on parallel machine vision algorithms and cutting-edge machine learning-based algorithms. Specifically, we are motivated to perform neural networks to extract higher-level helpful information from the analogue signals. An edge computing platform can be established based on our neural network with the PPA, where only a small quantity of extracted information is obtained, allowing for more efficient data transmission with less bandwidth. Hence, this thesis presents a lightweight and high-speed binary convolutional neural network on the sensor to categorise a range of objects. With the proposed methods to implement networks, all floating-point time-consuming multiplication operations can be replaced by efficient addition/subtraction and bit shifting operations. The focal-plane visual inference is difficult due to hardware resource constraints, such as limited registers and analogue noises. Hence, this work further proposes the purely binarised convolutional neural networks with both binary weights and activations. This thesis trains and implements neural networks with batch normalisation and adaptive threshold to binarise activations. The binary activations on the sensor benefit the neural network performance by alleviating the noises introduced by using analogue signals.

Hydraulic fracturing has attracted a tremendous amount of attention due to its abundant applications, such as magma-driven dykes, fracturing of oil and gas reservoirs, heat production from geothermal reservoirs, ice melt, and fault reactivation in mining. Therefore, the development of fracture propagation models is of great significance to provide more effective technical tools for the dynamic simulation of fracture formation and propagation. Hydraulic fracturing involves a complex process due to the strong coupling between the fluid and solid rock. The finite element method (FEM) is a traditional method for solving the fracture propagation problem. FEM has the advantage of being robust and flexible. However, special treatment is required when dealing with complex fracture problems such as fracture intersection and branching because of discontinuities across fracture surfaces. Instead of seeking to approximate partial differential equations in a conventional framework, the peridynamic approach (PD) uses spatial integral equations that overcome the limitations of defining partial derivatives at discontinuities, which makes it quite suitable for solving complex fracture problems. Therefore, it is of great importance to combine the advantages of both FEM and PD to develop a more robust and flexible fracture propagation model. The main contribution of the thesis is to study fracture propagation problems in fully saturated porous media based on both FEM and PD. Three fracture models based on FEM, PD, hybrid FEM and PD are developed to simulate and explore more "realistic" fracture propagation processes. The main research work of the thesis contains the following four parts. First, comprehensive reviews of the fracture propagation literature, crack branching and PD theory are given. The fracture propagation review includes a review of the fracture propagation process, fracture models and various numerical methods for predicting fracture propagation. The review of crack branching provides a state-of-the-art review of crack branching, including experimental observations, physics, fracture models and associated numerical methods. The latest advances and existing issues of crack branching are discussed. The review of PD theory provides a review of theoretical aspects of PD and related applica-tions of PD in different fields, especially in the field of fracture mechanics. The challenges and new prospects for the development of PD are discussed. Then, a fully implicit, FEM-based fracture model using zero thickness cohesive interface elements is developed. The rock formation is considered as a fully saturated porous medium. The fracture surface is deemed to be permeable, the fracturing fluid is treated as an impressible Newtonian fluid, and the cohesive zone model is employed for the fracture propagation criterion. After verifying the model, the stepwise phenomenon, which has been observed in the field and experiments and reported recently in the numerical modelling literature, is investigated with the model. Then, an explicit PD approach for simulating fracture propagation in saturated porous media is developed. A staggered method is adopted to solve the coupled system. The solid deformation equation is solved for the displacement field, and the fluid flow equation is solved for the pressure field with PD. After verification via benchmark examples, including a 1D consolidation problem and a 2D fluid-filled crack propagation problem, a series of tests are conducted to study the fracture propagation phenomena in porous media, including the crack branching phenomenon and the stepwise phenomenon. Effects of pore pressure in fracture propagation branching are investigated, and the stepwise phenomenon is successfully reproduced. All the examples presented demonstrate the capability of the approach in solving fracture propagation problems in saturated porous media. Finally, a fully coupled hybrid FEM-PD approach for simulating fracture propagation in saturated porous media is presented. Considering the ability of PD in solving discontinuous problems, the area where cracks can potentially occur is discretised by PD and the crack-free area is discretised by FEM. The solid deformation and fracture propagation are captured by PD and FEM, while the fluid flow in both the reservoir and fracture is simulated with FEM. The whole process is solved in a monolithic way with an implicit scheme. The approach demonstrates the capability of modelling complex dynamic crack propagation via benchmark examples. The branching phenomenon is then investigated with the proposed approach. In conclusion, complex crack patterns are prone to form with faster loading rate, more brittle and impermeable media, and with lower energy release rate.

This thesis is motivated by the need to reduce train aerodynamic noise, especially the noise from pantographs and bogies which are comprised of bluff bodies including cuboids. The underlying physics of the flow over cuboids with different cross-sections and the emitted noise are studied. The effects of wall proximity and corner radii on the flow field and noise of a cube as well as the effect of the aspect ratio of cuboids are the main focus. The Delayed Detached Eddy Simulation method is adopted to investigate the flow features and obtain the noise source information. The far-field noise is then predicted using the Ffowcs Williams-Hawkings acoustic analogy. Benchmark comparisons on the flow over a wall-mounted cube and a sphere are conducted to validate the feasibility of the adopted methodology for describing the flow field; a separate simulation on a square cylinder is also implemented to gain confidence in predicting surface pressure fluctuations and the far-field noise. Good agreement has been achieved between results from the benchmark cases and those from the literature. It is found that the wall proximity mainly affects the behaviour of the separated flow from the lower leading edge of the cube and also the flow separation along the ground. Large pressure fluctuations are observed close to the trailing edge and locations of the flow reattachment on the bottom of the cube. With the increase of the elevation height, the sound from the cube increases rapidly in the vertical direction and the maximum sound is observed when the cube is elevated by one quarter of the edge length above the ground. For rounded cubes, the flow separation is delayed with the increase of the corner radius; the shear layer moves towards the lateral surfaces and the recirculation length in the wake reduces. Flow features of the rounded cube vary with the Reynolds number, although no systematic dependence is observed. Large pressure fluctuations are found in the recirculating region on the lateral surface and near the trailing edge due to reduced recirculation length in the wake. Rounding the cube with a suitable radius can be an effective way to reduce the emitted noise. The minimum noise is found for a radius one third of the edge length. Cuboids with different aspect ratios are studied, where the aspect ratio is defined as the ratio of the width in the spanwise direction to the length in the streamwise direction. The flow separated from the upper leading edge only reattaches to the top surface for aspect ratios smaller than 1 whereas the vortex shedding in the wake tends to be more evident as the aspect ratio is increased. The Reynolds number has negligible effect on the mean flow behaviours of the cuboid. Large pressure fluctuations happen with the flow reattachment and the formation of the secondary vortex near the trailing edge. The emitted noise from the cuboid is always higher in the cross-flow direction than that in the streamwise and spanwise directions. The maximum sound is found for the aspect ratio of 6.

Many drugs have poor aqueous solubility which causes problems for their formulation as effective medicines. One way of solving this problem involves solubilizing drugs using aqueous dispersions of self-assembled surfactant micelles. It is well-established that surfactant micelles can solubilize drug, but where the drug will locate within the micelles and the factors that affect this are not fully understood. It is the locus of drug solubilization, however, that is thought to be important in determining the loading-capacity of the micelles, and it is this too that affects the drug's stability and its rate of release inside the body and so it also impacts the drug's therapeutic activity. In order to gain a better understanding of the location of drug in surfactant micelles, a series of systematic studies have been performed to determine the extent and locus of solubilization of two poorly water-soluble steroids, 4-cholesten-3-one (4-CHOL) and adrenosterone (ADRENO) compounds in a range of C12-chain surfactant of varying head group. The solubilisation capacities of steroids in the various surfactant micelle, as well as the interfacial and solvation behaviour of the various surfactant systems were explored. In order to estimate the preferred site(s) of solubilization of the steroids within the micelles, neutron techniques (both small-angle neutron scattering (SANS) and specular neutron reflectivity (SNR)) in combination with isotopic (H/D) contrast variation were used. The NR study revealed that regardless of the nature of the head groups, steroids were solubilised within the hydrophobic chain of surfactant at the air-water interface. The SANS studies were in good agreement with the NR study, and also suggested in some cases the presence of steroids could change the shape of the micelle. The information gained in these studies has provided for a more complete understanding of the relationship between the structure of a surfactant and the structure of the micelles likely to be formed when loaded with steroid drugs, allowing predictions to be made to guide the design of new surfactants to optimize steroid drug loading.

This thesis aims to explain the financial behavior of households and individuals in China, with a focus on the effects of the old-age dependency, household composition and healthcare systems. First, we investigate the association between old-age dependency ratio and household savings with 1995-2015 provincial-level panel data in China. The results show a negative association between the old-age dependency and the savings ratio, which is weaker in areas with higher level of government medical expenditure, financial development and insurance density. Second, we examine household composition and consumption with the 2011 and 2013 waves of China Health and Retirement Longitudinal Study (CHARLS). We provide evidence that the reallocation of resources freed up when an offspring moves out depends on the lever’s age. Finally, using the 2011, 2013 and 2015 waves of CHARLS, we evaluate and compare the Urban Employee Basic Medical Insurance (UEBMI) and resident health insurance (RHI) schemes. Estimations show that UEBMI is associated with a higher level of household consumption, the utilization of healthcare services, and medical expenditure (compared to RHI). Additionally, RHI fails to help poor people in purchasing sufficient healthcare, whilst UEBMI encourages rich people to overuse healthcare services.

The geopolitical conditions which led to the creation of NATO in 1949 rapidly disappeared following the fall of the Berlin Wall in 1989. The possibility of the termination of institutionalised US support for European security was seriously raised in this period, as was the possibility of NATO ceasing to exist. However, after progressive transformation, NATO expanded rather than disbanded. It went on to participate in ‘out of area’ actions. All these commitments were accompanied by debates about the purpose of NATO. Relevant debates included continuing tensions between Washington and European capitals over defence spending levels; accusations that the US was using NATO as an instrument of extra-United Nations unilateral power; the preference of Washington immediately after 9/11 for working through ad hoc rather than institutionalised alliance structures; and the developing relationship between NATO and Russia. This research seeks to explain the continuing US commitment to NATO in the post-Cold War era. The initial focus is on the recommitment decisions of the Clinton administration. It also researches in some depth the operations in Kosovo and, in particular, Libya. The case study on Libya is especially important in exploring the Obama administration’s understanding of the purpose of NATO in the context of current economic pressures, domestic US debates about post-War on Terror interventions, and of increasing American preoccupation with Pacific rather than European security. In the light of NATO operations in the post-Cold War era and the recent Syrian and Ukrainian crises, the study argues that the US has always been committed to NATO due to the unique value of NATO; the US overall foreign policy preference for NATO; and internal bureaucratic compromise on NATO. But the US may suspend its support to the Alliance in the future if the inherent problem of burden-sharing is not seriously treated by the European members.

At the outset of my PhD, my aim was to investigate the mechanisms responsible for the directed migration of primitive myeloid cells (PMCs) to wounds in Xenopus embryos. PMCs are the first blood cells to differentiate and become functional in Xenopus embryos, and have a notable migratory ability to be recruited by embryonic wounds before a functional vasculature is established. To find the mechanism underlying PMCs migration toward embryonic wounds, I first performed a screen to identify candidate cytoskeleton related genes, which might be responsible for facilitating the inflammatory response to injury in embryos. In situ hybridization and RT-PCR showed that coronin1a and l-plastin were specifically expressed in PMCs. I carried out loss-of-function experiments for coronin 1a and l-plastin in Xenopus embryos. Unfortunately neither knockdown affected the ability for PMCs to migrate during embryonic development or during the wound healing process. Loss-of-function experiments on coronin 1a and l-plastin also did not affect epidermal wound closure speed. Thus, although coronin 1a and l- plastin are expressed specifically in PMCs, they do not appear to be necessary for the migration of PMCs during development and during wound healing in Xenopuos embryos. Since my initial aim failed to provide insight into the mechanisms that mediate 9the inflammatory response to embryonic wounds, I decided to investigate the function of a previously identified monomeric actin protein during embryonic wound healing and appendage regeneration: namely Thymosin beta4 (Tβ4). In situ hybridization experiments showed that Tβ4 is expressed exclusively in the epidermis of developing frog embryos. Tβ4 knockdown embryos resulted in a significantly delay in the speed of wound closure during the early phase of wound healing. This delay correlated with a decrease in the actin contractile ring at the wound margin. Furthermore I found that the cell shapes of epidermal cells in the Tβ4 knockdown embryos were different from epidermal cells in control embryos. I hypothesize that this reduction caused the actin filaments changes in the epidermal cells, and were responsible for the failure of the cells to form an actin contractile ring, thus delaying the initial speed of wound closure. I tried to confirm that most of these defects specific to Tβ4, by performing rescue experiments with Tβ4 mRNA injections. Furthermore, I discovered that Tβ4 knockdown embryos displayed defects in tail development, including the absence of blood vessel branching within the fin of the tail. Finally, I found that the tails in Tβ4 knocked-down tadpoles failed to regenerate, while tails in control embryos regenerated completely following amputation. Both in situ hybridization and real-time PCR showed that Tβ4 was up regulated in the regenerated part of the tail in Xenopus tadpoles. Together with the tail amputation results, Tβ4 might be important for tail development and regeneration. These findings suggest that Tβ4 might play an important roles in the modulation of the actin cytoskeleton, which are essential for the proper behavior of epidermal cells during wound healing and appendage regeneration.

Environmental genotoxicants/carcinogens in the aquatic environment induce complex detrimental impacts on aquatic organisms. Integrated approach is required to comprehensively assess such impacts in terms of animal and environmental health. Biomarkers at different levels of biological organisations, including clearance rates (at individual level), histopathological alterations (at tissue level), DNA strand breaks using comet assay (at cellular level) and transcriptional alterations of key genes determined by real-time quantitative PCR (at molecular level), were designed and applied to analyse the impacts in marine bivalve: Mytilus sp., a classic bio-indicator species for environmental health monitoring, following exposure to selected environmental contaminants (i.e. Benzo(α)Pyrene (B(α)P), C60 fullerene, and tritium). The performance of biomarkers changed in response to exposures with higher sensitivity for DNA strand breaks and transcriptional alterations of selected genes. In particular, genes representing 7 gene groups (tumour suppressor gene - p53, proto-oncogene - ras, cytochrome P450 family - cyp4y1, MAPKs family - jnk, Bcl-2 family - bax, MDM2 - mdm2 and RAD family - rad51) showed tissue- and chemical-specific expression profiles under selected experimental conditions. A comparison revealed similarity of interplay among these highly conserved genes with mammalian models, suggesting Mytilus sp. could be an ideal model for signalling research in the carcinogenesis process. To my knowledge, this is the first research to detect tissue specific transcriptional alteration of p53 and ras genes in the organism for the analysis of detrimental impacts induced by different environmental genotoxicant/carcinogen exposures. Furthermore, the study is also the first one to include genes which can interplay with each other to function in the regulation of DNA repair, apoptosis and cell cycle arrest in marine organisms. The results suggested that the applied integrated biomarker approach is a comprehensive and sensitive method to monitor environmental health and, marine bivalve Mytilus sp. is an ideal model organism for research on genotoxicant induced carcinogenesis and will shed light on the mechanisms of this complex process in higher organisms, including humans.

This research investigates the relationship between political institutions and economic development and highlights the impact of political institutions on the economic development in China from 1992 to 2004 by discussing the reasons why the present political institutions in China require change. It is argued that political institutions could influence economic progress if the effect and role of political institutions - both positive and negative – is recognised. Two case studies are employed as examples of how the absence of appropriate political institutions affects the positive performance of the economy and how this is associated with China‘s history and the way in which economic reforms have been conducted. It is concluded that under the political monopoly of the single party, economic progress and development will be blocked and hijacked by the authorities and interest groups in China. Since its reform and opening up to the outside world, China‘s economy has so far seen an enormous growth, but these achievements are impressive merely in the short-term, and give a false impression of the economy‘s development. The free market economy system requires political reform but the Chinese Communist Party monopolizes all social resources and engages only in economic reform without political reform. This is what could be termed the “curse of the latecomer”: the long-term interests of the nation have been sacrificed and this may result in many hidden risks or even the failure of long-term development. This research identifies the major factors affecting the development of a country‘s society and economy. Political science theories about property rights and the State, and Institutional Change of New Institutional Economics are used to explain and support the standpoint of this thesis. Two case studies will be used in order to show how these theories are occurring in practice, which are the incidents concerning Yang Rong and Sun Dawu. The former will prove that it is necessary for property rights to be specified and enforced, and it is harmful to economic development when the government uses its political power to intervene in the economy. The latter case study will illustrate that the unfair monopoly and interventionist behaviour of government and a relatively defective legal system are not apt in facilitating the performance of China‘s economy. The conclusions of this study stress that institutions are the determinant of economic performance and that institutional changes are likely to occur when the existing institutions fail to satisfy people‘s demands. Such a development appears essential for China to progress further.