Abstract:Tuberculosis infection is one of the leading causes of mortality worldwide and is caused by Mycobacterium tuberculosis (Mtb). With an upsurge of multidrug-resistant tuberculosis, it is a global threat. Therefore, development of new drugs need immediate attention, and this needs identification of potential drug targets. The cell envelope of mycobacteria is one such attractive drug target owing to its role in maintaining the structural integrity and pathogenicity of the bacterium. The LytR-CpsA-Psr (LCP) family of proteins in Mycobacterium spp. have been shown to catalyze the coupling of arabinogalactan and peptidoglycan and possess pyrophosphatase activity. The four LCP protein homologues present in Mycobacterium smegmatis (Msmeg), MSMEG_0107, MSMEG_1824, MSMEG_5775 and MSMEG_6421, have not been extensively investigated with the focus on the existence and interplay of multiple LCP proteins. In this study with this non-pathogenic model organism, all four LCP homologues were shown to possess pyrophosphatase activity, with a significant higher activity displayed by MSMEG_0107 and MSMEG_5775. In order to further study the role of the LCP proteins on the physiology of the bacterium, single and double deletion strains lacking of the three non-essential lcp genes were created along with the respective complemented strains. All the generated mutants showed different phenotypes in the different assays, but usually not very severe. However, the double-deletion lcp mutant, ΔΔ(0107+5775) was the most affected mutant strain and displayed a disrupted cell envelope as evident from deprived growth rate, slower cellular aggregation, diminished biofilm formation on air-liquid interface, altered morphology, as well as an increased susceptibility to surface detergent, lysozyme and a wide range of antibiotics. Thus, the loss of both MSMEG_0107 and MSMEG_5775 exhibited profound effects on the mycobacterial cell envelope, and therefore could be further investigated as a possible combined drug target by extending these studies in Mtb. A novel approach in this study is the detection of exposed mycobacterial Galf moieties of arabinogalactan by EB-A2 monoclonal antibody, in the double lcp deletion mutant ΔΔ(0107+5775). Transcription profiling of all the lcp genes in the wild type strain and the mutants exhibited differential expression of these genes under both standard and stress conditions. A loss of MSMEG_5775 resulted in an upregulation of the other three lcp genes in comparison to the wild type strain under standard conditions. Under both acid and lysozyme stress, the loss of MSMEG_5775 downregulated all other lcp genes while loss of MSMEG_6421 upregulated these genes. Lastly, an in silico approach led to the identification of putative transcriptional factors in mycobacteria and related species which could be further investigated and experimentally confirmed. This study helped to understand the role of the lcp homologues in Msmeg better. From the differential expression studies, role of regulator(s) might be a significant approach to understand this family of proteins much better.

Abstract:With the increased demand for developing energy storage technologies, lithium-ion batteries have been considered as one of the most promising candidates due to its high energy density, excellent cyclic performance, and environmental benignity. Indeed, extensive applications of lithium-ion batteries are witnessed in the market, for example, in portable electronic equipment. However, the commercialized graphite anodes for lithium-ion batteries exhibiting low theoretical specific capacity is far from meeting the tremendous demands created by the fast-growing market. Therefore, enormous efforts have been devoted to developing desirable electrode materials with better recyclability and advanced capacity for next-generation lithium-ion batteries. Although alloy anode materials like silicon have the highest gravimetric and volumetric capacity, its huge volume change and low electron and ions conductivity still hinder the broad application in other fields, such as large-scale energy storage systems. Similar challenges also impede the wide implementation of conversion materials in Li-ion batteries. This work is focused on employing different highly conductive materials to improve the electrical conductivity of the entire electrode. At the same time, the formation of the conductive framework is beneficial to accommodate the substantial volume change of the active materials. In Chapter 3, copper nanowires and multi-wall carbon nanotubes coated on the surface of Cu foils built a porous substrate to support the active materials. Silicon was deposited on the porous substrate by the template of copper nanowires and multi-wall carbon nanotubes. The formation of copper nanowires/silicon and multi-wall carbon nanotubes/silicon core-shell structures intrinsically reduces the volume expansion of active materials. Meanwhile, the poles created by the intertwined copper nanowires and multi-wall carbon nanotubes further accommodate the stress from volume change. In addition, the copper nanowires/silicon and multi-wall carbon nanotubes/silicon core-shell structure provide the highly efficient electrons and Li+ diffusion pathways. As a result, we have demonstrated that multi-wall carbon nanotubes/copper nanowires/silicon delivers a high specific capacity of 1845 mAh g-1 in a half cell at a current density of 3.5 A g-1 after 180 cycles with a capacity retention of 85.1 %. In Chapter 4, a free-standing silicon-based anode was developed by preparing a three-dimensional copper nanowires/silicon nanoparticles@carbon composite using freeze-drying. Silicon nanoparticles were uniformly attached along with the copper nanowires, which was reinforced by the carbon coatings. The three-dimensional conductive structure allows the silicon nanoparticles to distribute evenly as well as enhance the electrical and ionic conductivity of the whole electrode. Similarly, considerable interspace produced by the three-dimensional structure can relieve the stress produced by the vast volume expansion of silicon nanoparticles, which is also restricted by the carbon coating layers during the charge and discharge processes. Moreover, the outer layers strengthen the stability of the three-dimensional framework and the contact between the copper nanowires and silicon nanoparticles. The electrochemical performance of copper nanowires/silicon nanoparticles@carbon composite electrode has been measured, which exhibits excellent cycling performance. In Chapter 5, a new highly conductive material, MXene nanosheets, was introduced to promote the electrochemical performance in lithium-ion batteries. In this chapter, the cobalt oxides were chosen as the active material for its controllable and facile synthesis methods. Meanwhile, cobalt oxides, one of the conversion materials as anodes for lithium-ion batteries, face similar issues with silicon. Therefore, an anode involving cobalt oxides nanoparticles mixed with MXene nanosheets on Ni foams has been developed. Small-size cobalt oxides nanoparticles were uniformly distributed within the MXene nanosheets leading to high lithium ions and electrons transmission efficiency, as well as preventing restacking of MXene nanosheets and colossal volume change of the cobalt oxides nanoparticles. As shown in Chapter 5, cobalt oxides /MXene composite electrode remains a stable capacity of 307 mAh g-1 after 1000 cycles when the current density approaches 5 C, which indicates the enormous potential of cobalt oxides/MXene composite as an anode for the high-performance lithium-ion batteries.

Abstract:This thesis studies the characteristics of Chinese futures markets and the quantitative investment strategies. The main objective of this thesis is to provide a comprehensive analysis on the performance of quantitative investment strategies in the Chinese market. Furthermore, with an econometric analysis, the stylised facts of the Chinese futures markets are documented. Extensive backtesting results on the performance of momentum, reversal and pairs trading type strategies are provided. In the case of pairs trading type strategies, risk and return relationship is characterised by the length of the maximum holding periods, and thus re ected in the maximum drawdown risk. In line with the increasing holding periods, the pro tability of pairs trading increases over longer holding periods. Therefore, the abnormal returns from pairs trading in the Chinese futures market do not necessarily re ect market ine ciency. Momentum and reversal strategies are compared by employing both high- and low-frequency time series with precise estimation of transaction costs. The comparison of momentum and reversal investment strategies at the intra- and inter-day scales displays that the portfolio rebalancing frequency signi cantly impacts the pro tability of such strategies. Complementarily, the excess returns of inter-day momentum trading with the inclusion of precise estimates of transaction costs re ect that quantitative investment strategies consistently produce abnormal pro ts in the Chinese commodity futures markets. However, from a risk-adjusted view, the returns are obtained only by bearing additional drawdown risks. Finally, this thesis suggests that investor should choose quantitative trading strategies according to the investment horizon, tolerance for maximum drawdown and portfolio rebalancing costs.

Abstract:Multi-carrier techniques play an important role in the fifth generation (5G) and beyond 5G (B5G) wireless communication systems, as they can support high data rate communications and exhibit high resilience to frequency selective fading. However, the presence of radio frequency (RF) impairments, such as carrier frequency offset (CFO), in-phase/quadrature-phase (IQ) imbalance, hinder the effectiveness of multi-carrier techniques. Thus, the estimation of RF impairments and channel are very essential. In this thesis, RF impairments(s) and channel(s), and their estimation together are considered for various multi-carrier 5G and B5G systems. This thesis consists of four main contributions as follows.  First, a joint multi-time of arrival (TOA) and multi-CFO estimation scheme is proposed for multi-user orthogonal frequency division multiplexing (OFDM) systems, where TOA is a key component of channel. With a carefully designed pilot, U TOAs and U CFOs of U users are separated jointly, dividing a complex 2U-dimensional estimation problem into 2U low-complexity one-dimensional estimation problems. Two CFO estimation approaches, including a low-complexity closed-form solution and a high-accuracy null-subcarrier assisted approach, are proposed to estimate the integer and fractional parts of each CFO as a whole. Each TOA estimate is robust against CFO by means of the features of the inter-carrier interference (ICI) matrix. Cramer-Rao lower bounds (CRLBs) of multi-TOA and mutli-CFO estimation are derived for multi-user OFDM systems. Extensive simulation results confirm the effectiveness of the proposed scheme.  Second, an iterative semi-blind (ISB) receiver structure is proposed for short-frame full-duplex (FD) OFDM systems with CFO. An equivalent system model with CFO included implicitly is first derived. A subspace-based blind channel estimation is proposed for the initial stage, followed by a single pilot assisted CFO estimation and channel ambiguities elimination. Then, channel and CFO are refined iteratively. The integer and fractional parts of CFO in the full range are extracted as a whole and in closed-form at each iteration. The proposed ISB receiver, with halved training overhead, demonstrates superior performances than the existing methods. CRLBs are derived to verify the effectiveness of the proposed receiver structure. It also demonstrates fast convergence speed.  Third, a robust semi-blind CFO and channel estimation scheme is proposed for generalised frequency division multiplexing (GFDM) systems. Based on an equivalent system model with CFO included implicitly, initial blind channel estimation is performed by subspace. Then, full-range CFO and channel ambiguity are estimated consecutively utilising a small number of nulls and pilots in a single subsymbol, respectively. Both CFO and channel estimates demonstrate high robustness against ICI and inter-symbol interference (ISI) caused by the nonorthogonal filters of GFDM. Simulation results verify that the bit error rate (BER) performance of the proposed scheme approaches the ideal case with perfect CFO and channel estimations.    Last but not least, a semi-blind joint estimation scheme of multiple channels, multiple CFOs and IQ imbalance is proposed for generalised frequency division multiple access (GFDMA) systems, with no constraints on carrier assignment scheme, modulation type, cyclic prefix length and symmetry of IQ imbalance. By means of subspace approach, CFOs and channels of U users are first separated into U groups. For each individual group, the CFO is estimated by minimising the smallest eigenvalue, whose corresponding eigenvector is utilised to determine channel. Then, IQ imbalance parameters and channel ambiguities are estimated jointly by very few pilots. Simulation results show that the proposed scheme significantly outperforms the existing methods, while at much lower training overhead. It also achieves a close performance to the derived CRLB.  To summarise, this thesis focuses on developing the estimation schemes of RF impairments and channels for 5G and B5G systems, by considering both OFDM and GFDM based multi-carrier techniques, half-duplex and full-duplex modes, single-user and multi-user systems. The developed estimation schemes are either pilot-aided with low complexity or semi-blind by subspace with high spectrum efficiency. This research work is an essential reference for academics and professionals involved in this topic. 

Abstract: The development of human activities has intensified and diversified the pressures applied to freshwater ecosystems. Particularly, land use stressors have been very pervasive and widespread. As a result, most freshwater systems are now under the influence of anthropogenic stressors. For instance, agricultural development and urbanisation have elevated the nutrients levels, facilitated the accumulation of chemicals, modified the natural flow velocities and promoted runoffs and sediment loads. Further, stressors often interact with each other, complicating the prediction of their effects on communities and ecosystem functioning; flow velocity and discharge reduction facilitate the accumulation of chemical and fine sediments. In order to evaluate the effect of multiple stressors and inform decision makers, investigations have been conducted worldwide on different trophic levels and ecosystem processes. Most notably, microbes, algae and macroinvertebrates have often been studied in isolation using taxonomic and now molecular methods. However, communities are made of complex population dynamics involving all trophic levels over time, and emergent ecosystem properties such as decomposition or net productivity are the result of multiple interactions between biotic and abiotic parameters. This calls for more holistic approaches encompassing as many facets of biodiversity as possible.   To investigate the effect of multiple land use stressors associated with agriculture and urbanisation, a highly replicated streamside field mesocosm experiment was built and performed in a near-pristine montane environment. The work was conducted in Autumn 2018 in the Jiulongfeng Nature Reserve, Huangshan, Anhui (China) and consisted of 64 experimental units naturally colonised by stream organisms for 3 weeks. I used a 4-factor full-factorial design, manipulating fine sediment deposition, flow velocity and nutrient concentration at two sampling times (2 and 3 weeks of exposure). Linear models were then applied to analyse the temporal response of microbial communities associated with both leaf litter decay and benthic biofilm formation, as well as the benthic macroinvertebrate communities. Additionally, to infer the emergent properties and functional characteristics of the different communities, four commonly used functional indices were investigated: (i) leaf litter decomposition in Chapter 2, (ii) databased predicted functional profile in Chapter 3, (iii) functional traits and (iv) functional diversity in Chapter 4. I then expanded my reflection from the knowledge acquired in the experimental side of my programme and outlined a novel framework to tackle multiple stressors interactions in riverine networks (Chapter 5).    The molecular analysis of microbial communities showed different impacts on species composition of the different stressors between microbes associated with leaf-litter decomposition and with biofilm development. Indeed, whilst nutrient enrichment and flow velocity reduction appeared to be the most pervasive factors affecting microbial decomposers communities on leaf substrates, fine sediment deposition and flow velocity reduction were most important for biofilm communities. Fine sediment deposition and flow velocity reduction were also the dominant factors driving macroinvertebrate community composition. Furthermore, both molecular analyses indicated that microbial clusters could be identified in response to the dominant stressors. In terms of interactions, 2-way interactions involving sediment and flow velocity reduction (sediment × flow velocity reduction) or nutrient enrichment and sediment (nutrient enrichment × sediment) were the most pervasive overall; 3-way interactions involving nutrient enrichment, sediment deposition and flow velocity reduction  (nutrient enrichment × sediment × flow velocity reduction) were also detected. Furthermore, temporal dynamics were also fairly widespread, highlighting the importance of integrating a temporal factor in multiple stressor studies. Finally, in accordance with the existing literature, changes in abiotic factors often led to functional rearrangements of the different communities underlying the environmental filtering and niche selection processes operating in the system.   From integrating the findings of this thesis into the wider subject area, I suggest ecosystem approach to multiple stressor interaction research. Specifically, I propose that future work adopt a spatiotemporal framework better integrating the energy fluxes across trophic levels and the flow of resources and material through riverine networks. Further, combining alpha diversity indices with functional traits aids understanding of the mechanisms that yield emergent ecosystem properties, such as productivity. Together, it is anticipated that spatiotemporal networks and functional measurements will facilitate prediction of the future stability of freshwater systems under stressor accumulation.  

Abstract:Lithium ion batteries (LIBs) have currently dominated the commercial market owing to the environmental benignity, suitable energy density, and long cycle lifetime. The commercial LIBs are commonly using graphite as anode materials, however, it has become clear that the theoretical capacity (~372 mAh g-1) of graphite has nearly reached the bottlenecks with little room for further exploration, and also the energy density and rate performance of existing LIBs are not sufficient for some advanced electronics equipment such as smart watch, and micro implantable biosensor system. With increasing demand and market potential, the worldwide academia researches and industrial community have been focused on investigating anode materials to achieve desirable power density, high rate performance, and long-term stability energy storage system, generating further impetus on flexible electrochemical applications, such as wearable devices, portable electronic devices especially for implant biological equipment. Alternative anode materials such as metal (Si, Ge and Sn) and metal oxide (Co3O4, SnO2 and GeO2) have been considered. Among them, the Si and germanium oxide have the highest theoretical gravimetric capacity in the elementary substance and oxide-based anode material respectively, which have been proposed as the best candidates for rechargeable battery anode. However, some challenges for these anode materials are also obvious due to the low conductivity and large volume expansion (> 300%) during the usage of LIBs. This expansion problem causes the pulverization of active materials and the repeated formation of the solid electrolyte interface (SEI) on that, resulting in the loss of interparticle electrical contact, and consequently deteriorating the battery cycle lifetime and capacity performance. In this work, we firstly demonstrated a facile method to fabricate a flexible alloyed copper/silicon core-shell nanoflowers structure anchored on the three-dimensional graphene foam as a current collector. In electrochemical testing, the resulting copper/silicon core-shell nanoflowered electrode demonstrates a high initial capacity of 1869 mAh g-1 at 1.6 A g-1, with a high retention rate of 66.6 % after 500 cycles. More importantly, at a high current density of 10 A g-1, this anode remains a high capacity retention > 63% (compared with the highest capacity 679 mAh g-1), offering enormous potential for energy storage applications. Secondly, we introduced a facile method to synthesize an amorphous GeOx-coated MXene nanosheet structure as the anode in lithium-ion batteries. For electrochemical performance, this GeOx/MXene nanosheet exhibited a reversible capacity of 950 mA h g-1 at 0.5 A g-1 after 100 cycles. It is indicated that the GeOx/MXene nanosheet structure can significantly improve the stability during the lithiation/delithiation prosses, with the enhanced capacity by the improvement of processes' kinetics. Thirdly, we built up a facile equipment to measure the high frequency capacitance change of silicon composite electrode. As this high frequency situation, the hypothesis circuit of the coin cell could be seemed as a combination of geometrical capacitance and resistance. For the alloy anodes which exhibited huge volume expanse during the lithiation/delithiation processes, the change of geometrical capacitance could be ascribed to the stress evolution and pulverization effect. Thereby the variation trend of the stress and pulverization could be determined by the change geometrical capacitance change.  To conclude, this project mainly focused on the pulverization and stress effect of the anode materials with alloying lithiation type. The strategies of first and second work were using the nanostructure engineering and 2D materials to release the stress and prevent the pulverization in the electrode. The results from these electrodes exhibited a stable electrochemical performance. Meanwhile, the rate performance of these electrodes was also improved by the additive of highly conductivity materials (e.g., copper, graphene, and MXene). To further investigate the consequence of severe volume expansion, we also built a high-frequency capacitance characterization system to perform the in-situ measurement of stress evolution and pulverization in coin cell with composite Si anode. That demonstrated the expected behavior corresponding to the electrode in the different states of charging.

Abstract:Solar photovoltaic (PV) energy is becoming an increasingly vital source in electricity grids for energy harvesting. Inspired by the regulatory incentives and plummeting cost, the integration of utility-scale PV systems into the power grid is boosting. Nonetheless, due to the natures of cloud movements, PV system exhibits rapid power ramp-rates (RR) in the output pro?les, which poses signi?cant challenges for system operators to maintain grid transient stability. In this context, this thesis focuses on the management of cloudinduced solar PV intermittency. Three aspects for coping with solar intermittency are addressed, namely, control, forecasting, and emulation. Firstly, from the control aspect, two predictive PV power RR control (PRRC) strategies are presented. To regulate system RRs, conventional methods are implemented either by active power curtailment (APC) or energy storage control (ESS). However, current APC method cannot deal with the ramp-down ?uctuations, and the integration of an ESS is still costly. On this point, two innovative PRRC strategies are proposed, which are based on a solar nowcasting system. The ?rst strategy does not require any ESS. With the prior knowledge of upcoming RRs, PV generation can be regulated before the actual shading occurs. The second strategy improves the conventional ESS method with minimal support of energy storage. The results show that both of the proposed strategies can e?ectively comply with RR regulations, and outperform the conventional methods. Then, in terms of forecasting, an improved sensor network-based spatio-temporal nowcasting method is developed. The proposed nowcasting method overcomes the shortcomings that typically associated with existing sensor network-based nowcasting methods, such as predictor mis-selection, inconsistent nowcasting, and poor model adaptability. The experimental results reveal that the proposed nowcasting method is more suitable for predicting system RRs. Subsequently, the operability of solar nowcasting for PRRC practice is demonstrated. To that end, temporal issues related to operational solar nowcasting are identi?ed, and their e?ects on nowcasting and PV control performance are evaluated. Lastly, from the emulation aspect, this thesis sets forth a partial shading emulator and a cloud shadow model, which can emulate the module-level responses of utility-scale PV systems under passing clouds. Based on the emulation tools, the characteristics of PV system RRs are comprehensively investigated across various system and cloud shadow attributions. The results indicate that a utility-scale PV system can frequently violate the RR limit imposed by grid operators. Hence, advanced RR control strategies should be essential for system operators to comply with the RR regulations.

Abstract:Wireless body and personal area networks have become commonplace in recent years in industrial, medical, and consumer-based applications, allowing a collection of devices such as medical sensors to be distributed around a person’s body or within their direct vicinity, to communicate with each other or a network controller to provide convenient personal services. Distributed devices are typically compact and can even be located within the human body. This produces several bottlenecks relating to RF ability and power availability which are addressed here. In this thesis, two antennas are developed. The first is designed for implantable and ingestible applications offering robust wideband performance, covering all the useable licenced operating bands, in the complex material characteristic environment of the human body. The radiation characteristics of the proposed antenna outperform other published work with a smaller size, achieved through the novel application of split-ring resonators. The second is an off-body antenna which concurrently provides appropriately polarised bands for indoor and outdoor localisation and data communication. For its minimised size and wide bandwidth, this antenna also outperforms other antennas for WPAN applications published in the literature. Two methods for simultaneous wireless information and power transfer have been proposed in this work, based on novel theoretical ideas and hardware implementations. A symbol splitting system separates the information- and non-information- carrying components of a signal, using each for data reception and energy harvesting, respectively. The second method makes use of the characteristic of the requisite rectifier in the power conversion from RF to DC, recycling the inevitable third harmonic for data reception. The hardware required to achieve both methodologies utilise couplers and each architecture has been proven feasible through simulation and measurement. They provide comparable performance to other published systems, offering a compact, efficient, and convenient route to simultaneous wireless information and power transfer.

Abstract:Discrete element method (DEM) is a useful numerical tool for analysing complex mechanical behaviour of granular materials as it considers the interaction at discrete contact points. In general, most of the DEM software packages use spherical particles by default because of easy contact detection and less computational cost. However, researchers confirmed that particle shape plays a significant role in exploring the mechanical behaviour of granular materials. Due to upgraded computation resources, nowadays it is possible to simulate the mechanical behaviour of granular materials considering true geometric shapes of particles. The key objective of the current research is to investigate the mechanical behaviour of granular materials considering particle shape characteristics. For that purpose, two basic geotechnical laboratory tests, i.e., direct shear test and triaxial test, are considered in this thesis.  The current research uses a commercial DEM code-named Particle Flow Code (PFC) developed by Itasca. An attempt was made to generate realistic particle shapes considering their major plane of orientations using a built-in clump mechanism in PFC. A series of DEM simulations were performed to investigate the sensitivity of the macroscopic specimen response to some specific parameter (e.g., particle numbers, loading rate). Based on the sensitivity analysis, selected microscopic parameters were selected to validate the DEM model with the experimental direct shear test results. To investigate the effects of particle elongations on the mechanical behaviour of granular materials, a series of simulations of direct shear tests and triaxial tests were performed using a range of dimensionless elongation parameters. The evolution of elongated particles was investigated at macro-and micro- scale levels. Moreover, the relationships between elongation parameter and critical state parameters were established.  A series of triaxial test simulations were performed considering two morphological descriptors and their mechanical behaviour was investigated at the macro- and micro-scale levels. In addition, a triaxial test environment was implemented to investigate the mechanical response of granular materials under different loading paths (i.e., axial compression (AC), axial extension (AE), lateral compression (LC) and lateral extension (LE)). The grain-scale interactions in terms of coordination number and deviator fabric were also investigated. Furthermore, the relationships were established among strength, dilatancy and state parameter concerning critical states. 

Abstract:Security is a significant problem for communications in many scenarios in Internet of Things (IoT), such as military applications, electronic payment, wireless reprogramming of smart devices and so on. To protect communications, a secret key shared by the communicating parties is often required. Authenticated key exchange (AKE) is one of the most widely used methods to provide two or more parties communicating over an open network with a shared secret key. It has been studied for many years. A large number of protocols are available by now. The majority of existing AKE protocols require the two communicating parties execute equivalent computational tasks. However, many communications take place between two devices with significantly different computational capabilities, such as a cloud center and a mobile terminal, a gateway and a sensor node, and so on. Most available AKE protocols do not perfectly match these scenarios. To further address the security problem in communications between parties with fairly unbalanced computational capabilities, this thesis studies AKE protocols with unbalanced computational requirements on the communicating parties. We firstly propose a method to unbalance computations in the Elliptic Curve Diffie-Hellman (ECDH) key exchange scheme. The resulting scheme is named as UECDH scheme. The method transfers one scalar multiplication from the computationally limited party to its more powerful communicating partner. It significantly reduces the computational burden on the limited party since scalar multiplication is the most time-consuming operation in the ECDH scheme. When applying the UECDH scheme to design AKE protocols, the biggest challenge is how to achieve authentication. Without authentication, two attacks (the man-in-the-middle attack and the impersonation attack) can be launched to the protocols. To achieve authentication, we introduce different measures that are suitable for a variety of use cases. Based on the authentication measures, we propose four suites of UECDH-based AKE protocols. The security of the protocols is discussed in detail. We also implement prototypes of these protocols and similar protocols in international standards including IEEE 802.15.6, Transport Layer Security (TLS) 1.3 and Bluetooth 5.0. Experiments are carried out to evaluate the performance. The results show that in the same experimental platform, the proposed protocols are more friendly to the party with limited computational capability, and have better performance than similar protocols in these international standards.

Abstract:The thesis contains three papers that focus on both private entrepreneurial firms and firms at the newly public stage. In regard to the privately entrepreneurial stage, Chapter 2 adopts an integrated signalling and screening perspective to investigate how investors perceive various signals sent by different firms across early financing stages. Through the use of multiple case studies of signaller?receiver dyads, Chapter 2 unexpectedly identifies a signal interpretation process model with three steps—extracting the fundamental signal; orchestrating signal compositions; and scrutinising signal consistency—and proposes differences among these three steps between the angel financing stage and the venture capital financing stage. Overall, Chapter 2 provides insights for the entrepreneurial financing literature by identifying a dynamic and temporal effect of different types of signals on high-technology entrepreneurial firms’ equity financing acquisition. Specifically, the findings of Chapter 2 indicate that some signals are persistent while others are temporary across different stages of a venture’s life cycle. Regarding the newly public stage, Chapter 3 examines how outside chief executive officer (CEO) succession affects newly public ventures’ growth. Building upon the evolutionary perspective, the thesis argues that outside CEOs can play a transformational role at the newly public stage because such CEOs are more aware of and motivated to break up organisational inertia for firm growth. The findings based on a sample of Chinese newly public firms between 2009 and 2018 indicate that newly public firms with outside CEO succession have stronger growth, and that this effect is stronger when these outside CEOs possess related experience in managing listed firms. The study further finds that following outside CEO succession, promotion of executives internally and adding of new senior executive roles can help outside CEOs to better play a transformational role, which strengthens firm growth. Based on the sample, Chapter 4 investigates the compromising decision arising from interactions between large shareholders based on the nature of different growth actions. Building on principal–principal agency theory, the thesis suggests that although acquisitive actions can promote growth, the second largest shareholder tends to discourage such growth action choices because of the potentially high agency risks. Instead, the second largest shareholder tends to encourage organic growth action choices even though such actions may produce a lower growth rate. The findings show that the second largest shareholder plays a dual role in monitoring the largest shareholder decision.

Abstract:RNA modification has emerged as an important layer for gene regulation, where biological functions are modulated by reversible post-transcriptional RNA modifications. N6-methyladenosine (m6A) is the most prevalent RNA modification on mRNAs and lncRNAs, and plays a pivotal role during various biological processes and disease pathogenesis. In this thesis, I presented the four bioinformatics approaches/applications to unravel the m6A epitranscriptome. We collected a total of 442,162 reliable m6A sites identified from seven base-resolution technologies and the quantified (rather than binary) epitranscriptome profiles estimated from 1,363 high-throughput sequencing samples to build the ‘m6A -Atlas’ database. As experimental approaches for studying the epitranscriptome are technically challenging and expensive, we used the collected base-resolution data to train a high-accuracy predictor ‘WHISTLE’ for m6A site identification from RNA sequences. Moreover, this prediction method was further extended to infer RNA modification-associated genetic variants to uncover potential epitranscriptome pathogenesis involving eight different types of RNA modification. In the last chapter, we presented a convenient measurement weighting strategy for enhanced detection of RNA co-methylation modules by tolerating the artifacts generated from epitranscriptome sequencing technology.

Abstract:Gonadotropin-releasing hormone (GnRH) is a pivotal regulator of the human reproductive system. Kisspeptin (KP) neurons act as the gate keeper of the GnRH neurons. GnRH and KP are both peptide hormones act on their cognate receptors, which are human GnRH receptor (hGnRHR) and KISS1 receptor (hKISS1R). They both belong to the superfamily of guanine nucleotide binding protein (G protein)-coupled receptors (GPCRs), which can activate G-protein dependent downstream signalling pathways, eventually lead to multiple cellular outcomes. In addition to their important roles in reproduction, they also inhibit proliferation and/or metastasis of cancer cells. Therefore, both of the receptors are important drug targets. NanoBRET-based ligand binding assays were developed to investigate ligand and receptor interactions. Fluorescently labelled peptide analogues, Fluorescent-GnRH and BODIPY630/650-GnRH, were firstly conceptualized, purchased, and evaluated. The Fluorescent-GnRH retained its receptor binding affinity, agonist activity and specificity, similar to that of the endogenous ligand, GnRH I. Thus, it was used to act as an energy acceptor in NanoBRET-based ligand binding assays, and also used in imaging studies of GnRHR. In addition, hGnRHR was tagged with Nluc at its N-terminus (N-Nluc-hGnRHR) to act as the energy donor in NanoBRET-based ligand binding assays. A secretory signal peptide (S) from interleukin-6 (IL-6) was added (S-N-Nluc-hGnRHR), to enhance its membrane expression. Similarly, a NanoBRET-based ligand binding assay with Fluorescent-KP-18 and S-N-Nluc-hKISS1R was also established. Overall,  a NanoBRET-based ligand binding assay format for high-throughput drug screening for hGnRHR has been established; thus, it has a great potential for drug development. Other NanoBRET-based assays were conducted to examine differential G protein coupling profiles of hGnRHR and hKISS1R, when stimulated with different ligands. The NanoBRET-based assay using hGnRHR-rTRHR tail-C-Nluc indicates that hGnRHR couples to Gi1, Gq, and G12, but not Gs, when stimulated with GnRH I and GnRH II. Overall, GnRH II seems to exhibit similar potency as GnRH I in activating Gq but seems to be less potent in activating Gi1 and G12. The reduction in activity is probably mainly caused by Tyr8 substitution in GnRH II. Similarly, after stimulation of KP-10 and KP-14, activated hKISS1R-C-Nluc showed a predominant coupling to Gq using the same assay. Additionally, activated hKISS1R-C-Nluc displayed a weak coupling to Gi1. However, whether this coupling is functional needs to be further investigated. To sum up, NanoBRET-based assays have been applied to determine ligand-induced selective signalling of hGnRHR and hKISS1R, and these findings could be important for the development of selective drugs that only activate desired receptor-mediated signalling pathways, while bypassing the others.

Abstract:This thesis presents research into the turbulent flow characteristic of open channels of complex cross-sections, with explicit attention to the interfacial region between the main channel and floodplain(s). The scope of the research includes two components: a series of detailed experimental investigations; the mathematical model development of coefficient of apparent shear stress for estimating zonal and overall discharge for complex asymmetric compound open channels. The primary goal of the experimental investigation was the procurement of high-quality data covering a well-defined and controlled range of hydrodynamic parameters. In support of the following objective, thirty-three sets of experiments have been undertaken and analysed in asymmetric compound channels to investigate flow behaviour over the floodplain and main channel interface. Laboratory experiments were performed under uniform flow conditions for new configurations with differential floodplain(s) width and multi-stage cases to fill the gap in the datasets of asymmetric compound open channels. Fundamental measurements were taken for estimating depth-averaged velocity, Reynolds shear stress, secondary currents, and apparent shear stress to examine the transverse current interaction of two-stage for new configurations of asymmetric compound channels using down-probe and side-probe acoustic Doppler velocimeter. The objective of the theoretical investigation was to obtain the generalized behaviour of the new configuration for the interaction mechanism. Flow interaction between two sub-sections affects the overall discharge capacity and conveyance distribution in compound open channels. Many investigators attempted to estimate flow interaction regarding apparent shear stress acting on the imaginary plane between the floodplain and main channel. However, previous models are neither generalized for asymmetric channels nor applied to a wide range of data sets, including field data, even though the apparent shear stress for asymmetric channels is higher than symmetric channels for the same flow depth and geometrical congruency. The momentum exchange models used in this thesis were motivated by scaling arguments and allowed a simple analytical solution for the zonal discharge in each section. However, it was found that the apparent shear models perform differently based on different depth ratios. None of the previous models performed well in channels with a low depth ratio. The different models for apparent shear based on width ratio and slope were found to give mixed results, discussed in detail. The resulting new models for the coefficient of the apparent shear stress are proposed to improve the zonal and overall discharge estimation for these new configurations. Models revealed that the coefficient is strongly dependent on the depth ratio for different ratios of bankfull height to floodplain width in these new configurations. The proposed new models can be applied to laboratory and field data without calibration.

Abstract:The modern smart grid requires flexible control ability, high transmission efficiency, and good robustness due to contingencies. Besides, a growing number of power stations and load is Direct Current (DC) power, such as photovoltaic power stations, battery energy storage stations, most consumer electronics like a computer. Thus the DC power transmission systems, such as DC solid-state transformers (SSTs) can be utilized to reduce the volume and losses of the transmission system. Among various DC SSTs structures, the DC SSTs based on dual active bridge (DAB) converter is considered as promising topology due to its symmetrical structures, bidirectional power flow capacity, wide soft switching region and flexible control ability. As a key component of DC-SSTs, the operation of the DAB converter will determine the overall performance of the whole system. Thus the improvement of DAB is essential to DC-SSTs and modern smart grid applications. In this thesis, steady-state and dynamic state operation, as well as soft switching behavior of DAB converter, have been studied. For improving the steady-state performance of the DAB converter, multiple optimizations are proposed to reduce the backflow current or reactive power and extend the soft switching region for improving the transmission efficiency. Besides, the frequency domain model is introduced to further reduce the complexity of the optimization model. The effectiveness of those optimization schemes has been verified by experimental results. Compared with traditional phase-shift control, these proposed optimization methods can significantly increase the transmission efficiency. Furthermore, the multiple natural switching surfaces boundary control is proposed to enhance the dynamic performance of the DAB converter, especially for start-up and voltage variation conditions. It can achieve a fast-dynamic response and eliminating DC bias current. Both simulation and experimental results have been presented to prove the superiority of the proposed method. Compared with traditional closed-loop control based on the PI controller, the proposed boundary control can dramatically accelerate the dynamic response. Moreover, the resonant transition for different switching conditions during the dead-time period has been investigated. Then, the phase correction method and variable dead-time are proposed to compensate the phase difference between the gate signal and actual waveform and power losses during dead-time. The effectiveness of those methods is validated by comparing the proposed method with a fixed dead-time method through the experimental result. It suggests that the proposed dead-time compensation and various dead-time method can correct the phase delay and improve transmission efficiency. 

Abstract:The primary objective of this research is to investigate various optimization problems connected with partial differential equations (PDE). In chapter 2, we utilize the tool of tangent cones from convex analysis to prove the existence and uniqueness of a minimization problem. Since the admissible set considered in chapter 2 is a suitable convex set in $L^infty(D)$, we can make use of tangent cones to derive the optimality condition for the problem. However, if we let the admissible set to be a rearrangement class generated by a general function (not a characteristic function), the method of tangent cones may not be applied. The central part of this research is Chapter 3, and it is conducted based on the foundation work mainly clarified by Geoffrey R. Burton with his collaborators near 90s, see [7, 8, 9, 10]. Usually, we consider a rearrangement class (a set comprising all rearrangements of a prescribed function) and then optimize some energy functional related to partial differential equations on this class or part of it. So, we call it rearrangement optimization problem (ROP). In recent years this area of research has become increasingly popular amongst mathematicians for several reasons. One reason is that many physical phenomena can be naturally formulated as ROPs. Another reason is that ROPs have natural links with other branches of mathematics such as geometry, free boundary problems, convex analysis, differential equations, and more. Lastly, such optimization problems also offer very challenging questions that are fascinating for researchers, see for example [2]. More specifically, Chapter 2 and Chapter 3 are prepared based on four papers [24, 40, 41, 42], mainly in collaboration with Behrouz Emamizadeh. Chapter 4 is inspired by [5]. In [5], the existence and uniqueness of solutions of various PDEs involving Radon measures are presented. In order to establish a connection between rearrangements and PDEs involving Radon measures, the author try to investigate a way to extend the notion of rearrangement of functions to rearrangement of Radon measures in Chapter 4.

Abstract:Clustering is one of the most widely used pattern recognition technologies for data analytics. Density-based clustering is a category of clustering methods which can find arbitrary shaped clusters. A well-known density-based clustering algorithm is Density- Based Spatial Clustering of Applications with Noise (DBSCAN). DBSCAN has three drawbacks: firstly, the parameters for DBSCAN are hard to set; secondly, the number of clusters cannot be controlled by the users; and thirdly, DBSCAN cannot directly be used as a classifier. With addressing the drawbacks of DBSCAN, a novel framework, Evolutionary and Swarm Algorithm optimised Density-based Clustering and Classification (ESA-DCC), is proposed. Evolutionary and Swarm Algorithm (ESA), has been applied in various different research fields regarding optimisation problems, including data analytics. Numerous categories of ESAs have been proposed, such as, Genetic Algorithms (GAs), Particle Swarm Optimization (PSO), Differential Evaluation (DE) and Artificial Bee Colony (ABC). In this thesis, ESA is used to search the best parameters of density-based clustering and classification in the ESA-DCC framework to address the first drawback of DBSCAN. As method to offset the second drawback, four types of fitness functions are defined to enable users to set the number of clusters as input. A supervised fitness function is defined to use the ESA-DCC as a classifier to address the third drawback. Four ESA- DCC methods, GA-DCC, PSO-DCC, DE-DCC and ABC-DCC, are developed. The performance of the ESA-DCC methods is compared with K-means and DBSCAN using ten datasets. The experimental results indicate that the proposed ESA-DCC methods can find the optimised parameters in both supervised and unsupervised contexts. The proposed methods are applied in a product recommender system and image segmentation cases.

Abstract:As a lesson from the severe losses of $827 million by UK merchant bank Barings in 1995 , qualitative and quantitative modelling in operational risk started to attract more research attention in banking and insurance system. Such an operational risk would lead to serious consequence, even bankruptcy. Hence, it is necessary for financial institutions to model and avoid the operational risk. With this in mind, this thesis investigates important topics in quantitatively estimating of operational risk. We use heavy-tailed distribution functions to model the loss severities, and use several tools, such as copulas and multivariate regular variation, to model the dependence structures. Firstly, we consider both univariate and multivariate operational risk models, in which the loss severities are modelled by a series of weakly tail dependent and heavy-tailed positive random variables, and the loss frequency processes are general counting processes. In such models, we study the limit behaviors for the Value-at-Risk and Conditional Tail Expectation of aggregate operational risks. The methodology is based on capital approximation within the framework of the Basel II/III regulatory capital accords, which is the so-called Loss Distribution Approach. We also conduct simulation studies to check the accuracy of our obtained approximations and the (in)sensitivity due to different dependence structures or the heavy-tailedness of the severities. Next, in order to include both the weakly and strongly tail dependent case, we first consider a bivariate operational risk cell model, in which the loss severities are modelled by heavy-tailed and weakly (or strongly) dependent nonnegative random variables, and the frequency processes are described by two arbitrarily dependent general counting processes. In such a model, we then establish asymptotic formulas for the VaR and CTE of the total aggregate loss. Simulation studies are also conducted to check the accuracy of the obtained theoretical results via the Monte Carlo method. Later, we further extend the study of Gong and Yang (2021) to derive asymptotic approximations for VaR and CTE in a multivaraite operational risk cell model. Consider a multivariate operational risk cell model, in which the loss severities are modelled by heavytailed and weakly (or strongly) dependent nonnegative random variables, and the frequency processes are described by several arbitrarily dependent general counting processes. In this model, we establish asymptotic formulas for the VaR and CTE of the total aggregate loss. Numerical studies are conducted to examine the performance and to test the sensitivity of these asymptotic formulas. Lastly, to study the randomly weighted sums with infinitely many dependent terms, we consider the randomly weighted sums generated by a series of dependent subexponential primary random variables and a few arbitrarily dependent random weights. Then, we establish a Kesten-type upper bound for their tail probabilities in presence of subexponential primary random variables and under a certain dependence among them. As applications, then we derive asymptotic formulas for the tail probability and the VaR of total aggregate loss in a multivariate operational risk cell model.

Abstract:Vegetation plays a fundamental role in changing the flow characteristics of natural channels such as rivers. There are some studies which evaluated the impact of vegetation in open channel flows, but their modelling are not close to nature situation and usually only cover on single type of vegetation. Vegetation in natural channels is usually denser in lower layer and sparser in upper layer. For example, in riparian environments or floodplains, shorter vegetation (grasses or shrubs) is submerged, but the taller vegetation (e.g. trees) remains emergent. However, the impact of such mixed layer vegetation on the flow structure is not understood well, which is significant to reduce risk of flood and water environment. In this thesis, a series of laboratory experiments have been undertaken to study the impact of double layer vegetation in both emergent and submerged conditions. The vegetation was simulated by an array of PVC dowels with two different heights of 10 cm and 20 cm. The experiments were carried out in a rectangular hydraulic flume in Nanjing Hydraulic Research Institute and Xi’an Jiaotong-Liverpool University, respectively. Dowels were arranged in 5 different formations with each having 4 different flow depths to capture inflection of velocity over the mixing region between short and tall dowels. Velocity measurements were taken by using 3-D Acoustic Doppler Velocimeter (ADV) and Propeller Velocimeter in order to obtain key parameters such as turbulence intensity, Reynolds stress and turbulence kinetic energy. Ansys Fluent was used to simulate the same sets of vegetation configurations using K-? model with mesh sensitivity analysis to capture the inflection over the short vegetation region. The numerical study was explored for the double layer vegetation, and showed that the modelling results have good agreement with the experimental data for different vegetation configurations. New analytical models based on Reynolds-averaged closure principles have been also proposed to describe the vertical distribution of mean streamwise velocity in an open channel flow with double-layered vegetation. The proposed models were evaluated with extensive experimental data from our experiments and other published experiments available in the literature. The Root Mean Square Error (RMSE) of the velocity comparisons is found to be less than 0.0342 m/s, which is acceptable. In another series of experiments, vegetation dowels have been located only in one side of the channel while the other side is empty to simulate the partial vegetation and it has been noticed that a strong shear layer exists between non-vegetation and vegetation zones, indicating the reduction effect of vegetation on the velocity of flow. Furthermore, modifications are recommended to properly calculate the hydraulic radius and Manning's coefficient for the flow with double-layer vegetation. Finally it has been concluded that the flow in double layer vegetation is more complicated compare to flow through single layer vegetation and therefore the calibration of proposed models with three layers vegetation has been recommended.