This thesis investigates the potential of aesthetics in the design of Human-Computer Interactions (HCI). In particular, it aims to provide a means by which aesthetics can be applied in photorealistic virtual reality (VR) to create ‘engaging' experiences. Indeed, the author of this thesis suggests that much can be gained from looking at the aesthetics of photorealistic VR content as opposed to the more traditional HCI approaches that have primarily focused on the performance and efficiency issues of the technology. The thesis is motivated by the very notion that the aesthetic potential of photorealistic VR content is, and continues to be, underestimated whilst the emphasis on the development of newer and more efficient visualisation technologies to create engaging VR experiences increases. Challenging this, the research is firmly based on the premise that the aesthetic content of photorealistic VR environments holds at least as many possibilities for the creation of more complete and ‘engaging' experiences. To explore this, the thesis describes the aesthetic-interaction as a new type of interaction that focuses primarily on how one aesthetically interacts with an interface (as opposed to how one cognitively interacts with interface metaphors to activate certain aspects of its functionality). By concentrating on the design and the aesthetic content of photorealistic VR, more so than the building and enhancing of its functional capabilities, the research aims to probe how the user might be sensually attracted and aroused into the sharing of information and hence the creation of new and ‘engaging' experiences. In particular, this thesis examines the role of narrative in VR and from this, moves to the visual-narrative which for centuries has employed the use of aesthetics to ‘engage' the spectator in its process of storytelling. Integrating traits of the VR medium and the visualnarrative, the author develops a visual-narrative structure which is used to elicit a number of design requirements specifically for photorealistic VR. From this, a visual-narrative model is developed, and in a second study, it is demonstrated that the photorealistic VR environments designed with the visual-narrative model can, indeed, create more ‘engaging' experiences. In summary, this research provides a means by which the aesthetics of photorealistic VR (or other HCI technologies) might be strategically patterned and applied to the creation of various different types of ‘engaging' experiences. By describing the aesthetic-interaction and how it can be articulated through a visual-narrative model, the research not only successfully highlights the experiential side of photorealistic VR, but also advances the new ‘design' drive of HCI.

This thesis describes research conducted into the measure- ment of pedestrian movement. It starts with an examination of current pedestrian detection and tracking systems, looking at several different technologies including image-processing systems. It highlights, as other authors have, that there is still a substantial gap between the abilities of existing pedestrian measurement and tracking systems and the requirements of users of such systems. After the review it provides an introduction to human gait and its use as a biometric. It then examines the IRISYS people counter, a low resolution infrared detector, used for this research. The detector's advantages and disadvantages are discussed, a detailed description of the data produced is provided. The thesis then describes in detail a study establishing that human gait information can be measured by the IRISYS people counter. It examines the use of the detectors in stereo to measure the height of the people; however the results are not impressive. During this investigation the presence of oscillations likely to relate to this walking gait is noted in the data. A second study is carried out confirming that the noted oscillation originates from human gait and further data is gathered to enable the development of measurement algorithms. The magnitude of the walking oscillation noted is examined in detail. It is found to be both individualistic and highly correlated to gender. A gender distribution algorithm is presented and evaluated on data captured in two different locations. These show very promising results. Several different methods are described for processing the infor-mation to extract a measure of cadence. The cadence is found to be individualistic and shows interesting correlations to height and leg length. This thesis advances the field of pedestrian measurement by conducting pedestrian motion studies and developing algorithms for measuring human gait.

Electronic Business (e-Business) is recognised as being able to help organisations to achieve competitiveness benefits. In Libya, the uptake of e-business in the higher education sector has been relatively limited and ineffective. There is no guidance available to help Libyan universities to maximise their leverage by participating intelligently in e-business deployment. This research examines e-business deployment initiatives and provides insights and learning into how to successfully develop and implement these initiatives within universities in a developing country, specifically Libya. The aim of this research is to develop a robust framework to support an efficient e-business deployment focusing on the case of Libya. This is achieved by exploring selected completed and on-going successful initiatives and focusing on the barriers to, and the enablers of, these initiatives. As a result, the nature of successful e-business deployment is determined, and the e-business deployment requirements for universities in Libya are identified. A combination of research methodology approaches has been engaged in this research. Firstly, an extensive review of literature took place to summarise and create the arguments around the main e-business models and their related dimensions, contributing to the development of e-business deployment research. Secondly, a qualitative approach based on case studies, include an explanatory investigation in three cases, is selected as an appropriate methodology for this research. Questionnaires and semi-structured interviews are conducted to gather data from selected universities. The results show that Process, People and Technology barriers are the main barriers facing e-business deployment in Libyan universities. The collected data was analysed and problems were identified; as a result, the elements for a holistic approach to manage e-business deployment emerged. Based on the findings, an e-business deployment model is developed to provide a holistic approach for universities to plan, perform and review their e-business levels, as well as uncovering the main barriers preventing these universities from expanding e-business development. Also, a model of the implementation process of e-business deployment is proposed. From a practical point of view, this research provides realistic implications for decision makers and officials within Libyan universities involved in the process of planning, developing and implementing e-business initiatives.

Learning and teaching have been influenced greatly by the rapid development of technology. For instance, through the use of soft computing techniques, it would be possible to create an artificially intelligent autonomous tutor agent, which can ease the burden on teachers and enhance learning outcomes through its more personalised interaction with students. Providing students with automated guidance, such as directing students through the most appropriate content sequence is one aim of online tutoring systems. However, in most of the available tutoring systems, users neither have the ability to adjust the tutor agent’s autonomy level nor fully control the rules applied by the tutor agent. Thus, this thesis has sought to overcome these shortcomings by proposing a system called the ‘Adaptive Course Sequencing Approach’ (ACSA) which enables students to adjust the autonomy level of the tutor agent and gives teachers the ability to directly communicate with the tutor agent to create the sequencing rules and alter them at any time during the learning experience. This is achieved with fuzzy logic, which has the capability of producing human-readable sequencing rules as well as managing the uncertainty of measuring some students’ levels of knowledge. We hypothesise that by equipping intelligent educational environments with adjustable autonomy mechanisms, the students’ learning outcomes will be enhanced. This research was divided into seven phases and involved a large number of participants (1725 in total) to assess the need for adjustable autonomy mechanisms in online tutoring systems and to explore the way of providing these mechanisms in ACSA, thereby demonstrating the hypothesis by two empirical experiments. The results showed that applying adjustable autonomy mechanisms significantly improved the students’ learning outcomes and that the students who adjusted the autonomy level more than once performed slightly better than those who adjusted it once only. In addition, applying the collaborative-driven agent method, which relies on machine learning to generate and optimise the sequencing rules, led to improving the students’ learning outcomes and highly satisfying the teachers.

The Web has changed the face of scientific communication; and the Semantic Web promises new ways of adding value to research material by making it more accessible to automatic discovery, linking, and analysis. Institutional repositories contain a wealth of information which could benefit from the application of this technology. In this thesis I describe the problems inherent in the informality of traditional repository metadata, and propose a data model based on the Semantic Web which will support more efficient use of this data, with the aim of streamlining scientific communication and promoting efficient use of institutional research output.

This report summarizes the work that has been carried out as part of the FLAVIIR project, a 5 year research program looking at technologies for future unmanned air vehicles. A novel classication of aircraft product defnition is utilised and a framework to estimate the life cycle cost of aircraft using the product definition is presented. The architecture to estimate the life cycle cost and the associated models are described. The acquisition costs are estimated using a hierarchical structure and a discrete simulation model is used to estimate the maintenance and operation costs. The acquisition cost model uses an object oriented approach with libraries of materials and processes integrated into the cost model. Risk analysis is performed to identify the important design parameters and uncertainty in the model. The acquisition cost model developed has the capability to estimate the costs of aircraft structures manufactured using metal-based materials as well as non-metal-based materials. The discrete event simulation model estimates the operation and maintenance costs of a fleet of aircraft using the mission characteristics, aircraftperformance and the logistics data as input. The aircraft performance parameters are calculated by using aerodynamic analysis along with performance analysis models and the simulation model utilises a novel methodology to link aircraft performance with survivability analysis for estimating the maintenance costs. A framework is presented in which the cost models developed can be integrated into the conceptual design process to facilitate the comparison between different configurations. The usage of the life cycle cost framework as a decision support tool is outlined and three case studies are presented which include composites vs metals trade-off analysis, optimisation studies and web deployment for real time cost estimation. The novel contributions of this research are outlined and interesting avenues for future research that can be pursued are identified.

This thesis explores the prediction of rearwards fan noise propagation within the bypass duct and its radiation into the far feld. Two recently developed models: B-induct and GXMunt, are exploited in application to real engine bypass ducts and their performance is evaluated. These methods are an improvement on current industry standards, allowing realistic duct geometry and flow conditions to be modelled with reasonable computation and time demands. The main focus is on the model b-induct. B-induct predictions for bypass attenuation are integrated into an industry standard whole engine model, and predictions of far-feld noise are obtained for a modern high bypass-ratio engine. These predictions compare more favourably with measured data from full-scale static engine tests than similar predictions made using a standard uni-form rectangular duct model for the bypass attenuations, indicating that b-induct is an improvement over the current model. Initial studies on the effect of duct geometry on noise propagation suggest a noise benefit for a duct with higher curvature when compared to a typical Baseline design. This suggestion is confirmed using measured data from zero-flow rig tests. Predictions for three-dimensional duct geometries are also performed to show the effect of scattering due to bifurcations within the duct. B-induct allows for the specific bypass geometry and liner positions to be taken into account when performing impedance optimisations. A new optimisation procedure is proposed in which b-induct predictions are used within an existing whole aircraft noise prediction model. This procedure is used to select liner impedances for a modern engine bypass duct. B-induct is demonstrated to be a promising new tool within the engine design process, for both analysis of the impact of rear fan noise on whole engine noise, and assessment of potential low noise bypass configurations.

In this work, a new thermo-mechanical model is developed by improving on an existing one, applicable to large deep seated landslides and rockslides consisting of a coherent mass sliding on a thin clayey layer. The considered time window is that of catastrophic acceleration, starting at incipient failure and ending a few seconds later, when the acquired displacement and velocity are such that the sliding material begins to break up into pieces. The model accounts for temperature rise in the slip zone due to the heat produced by friction, leading to thermoplastic collapse of the soil skeleton and subsequent increase of pore water pressure. This in turn drastically decreases the resistance to motion and allows the overlying mass to move downslope ever more freely. The proposed model is implemented numerically and validated by back-analysing the two well-documented catastrophic landslide case histories of Vajont and Jiufengershan. The model is then employed to carry out a parametric study to systematically investigate the development of catastrophic failure in uniform slopes. It was found that the most influential parameters in promoting catastrophic collapse are (1) the static friction-softening rate a1, (2) the slope inclination β, (3) the soil permeability kw, (4) the dynamic residual friction angle rd φ and (5) the overburden thickness H. The most dangerous situation is when a1, β and H are very large and kw and rd φ are very low. Of the above, the ‘thermo-mechanical parameters’ kw and H deserve more attention as they have been introduced by the thermo-mechanical model and are not normally considered in standard stability analyses of uniform slopes. A second parametric study was performed to demonstrate that thermo-mechanical parameters alone can make a difference between a relatively non-catastrophic event and a catastrophic one. Hence, further insight into the design of landslide risk mitigation measures can be gained if, in addition to the standard site investigations, the permeability of the soil is measured and the depth of an existing or expected failure surface is measured or estimated respectively.

The search for the most effective method for the geometric parameterization of many internal fluid flow applications is ongoing. This thesis focuses on providing a general purpose automated parameterization strategy for use in design optimization. Commercial Computer-Aided Design (CAD) software, Computational Fluid Dynamics (CFD) software and optimizer tools are brought together to offer a generic and practical solution. A multi-stage parameterization technique for three-dimensional surface manipulation is proposed. The first stage in the process defines the geometry in a global sense, allowing large scale freedom to produce a wide variety of shapes using only a small set of design variables. Invariably, optimization using a simplified global parameterization does not provide small scale detail required for an optimal solution of a complex geometry. Therefore, a second stage is used subsequently to fine-tune the geometry with respect to the objective function being optimized. By using Kriging response surface methodology to support the optimization studies, two diverse applications, a Formula One airbox and a human carotid artery bifurcation, can be concisely represented through a global parameterization followed by a local parameterization.

Magnetic exchange spring systems are multi-layers or composites of magnetically hard and soft materials that are exchange-coupled across their interfaces. In recent years, research into exchange spring systems has flourished, with potential for application in high-performance permanent magnets, GMR spin devices, magnetic MEMS technology, and in magnetic data storage. We investigate the magnetic properties of MBE grown superlattices with alternating layers of magnetically hard rare earth-iron (DyFe2, ErFe2) and soft yttriumiron (YFe2) compounds. They are ideal model systems to study exchange spring phenomena. We develop numerical models of the investigated systems and apply micromagnetic simulations. The simulation code OOMMF is extended and used to solve Landau-Lifshitz-Gilbert and Brown’s equations. This allows us to determine the microscopic configuration of the magnetisation that is not directly accessible by experiment. Magnetic field-sweep measurements of a multilayered DyFe2/YFe2 system show an unexpected triple switching of the magnetically hard DyFe2 layers. The magnetisation of the hard magnetic layers reverse before the soft magnetic layers. We reproduce the experimental hysteresis loops of the net and compound-specific magnetisation by means of simulations and explain the switching behaviour. Using similar numerical methods, we interpret experimental data on ErFe2/YFe2 multilayers. At sufficiently high fields, applied perpendicular to the multilayer film plane, the energy is minimised by a multilayer spin flop. This is a particular spin configuration where the magnetisation aligns with a direction perpendicular to the applied field. Taking the preceding findings further, we investigate multilayers of ErFe2/YFe2/ DyFe2/YFe2. We gain insight in the complex spin configurations in systems of different magnetically hard materials, with a pre-strung domain wall in the soft YFe2 layers. Varying the thickness of the YFe2 layers, we study the changing mutual interference of the switching patterns in the ErFe2 and DyFe2 layers.