Search

A common practice for solving problems in video analytics is to consider each problem (e.g. processing speed, complexity of data analysis or tracking of moving object) in isolation, starting with constructing of computational models for each specific data processing task. This is dictated by the need to address the complexity of the overall problem, which typically can be broken into multiple computational tasks - detection, recognition, classification, analysis, prediction, etc. Each of these tasks has its own computational model and uses different data. This approach has been the main focus of research in video processing for a long time and had some success, but it leads to the significant requirement for computational power and is not feasible for real-time video analytics tasks because of the time constraints. An alternative to this strategy is to link the video analytics tasks in order to leverage the related data sets and computational models. By doing that we can decrease the amount of data and computations necessary and thus, reach satisfactory performance. Consequently, we can improve the accuracy of the solution obtained by simplification through utilizing the previous experience through learning. Our aim in this research is to exploit the variety of parameters of this process - development of the data processing model, features extracted from the data and algorithms of the machine learning. This research addresses the problem of reconstructing moving object trajectories for real-time video analytics based on continuous processing of input video streams with application to a wide range of tasks in video surveillance. Our approach uses a combination of classifiers for different features of the data, such as shape and colour, as well as other parameters (e.g. the position of human body parts) which can be used for detecting various profiles of interest for further processing of the original video stream. The custom-tailored model of moving object is created by experimenting with different models and algorithms specifically applicable to the task for trajectory reconstruction and selected to make the best use of data and to achieve the best result in terms of speed of data processing which is further discussed in chapter 6. On the basis of the experimental work conducted in this research we have developed a software system, which can track and identify moving human objects of interest within live video streams through a combination of visual pre-processing (discussed in chapter 5 in video transformation module), machine learning and video post-processing (discussed in chapter 5 in ‘rotation’ activity) methods. It uses a novel object detection algorithm which significantly improves the speed of processing during the tracking and localization steps. This algorithm makes use of a number of comparing analytics method by classifying the shapes and sizes parameters as shown in chapter 4 with different cases. Although specifically designed to serve the purpose of the research programme for behaviour analysis of moving objects in live video streams this part of our framework can be considered as a standalone problem solving module for software systems which involve object detection, identification and tracking based on live video input. It can be used as an open source software in many areas ranging from public safety management and video surveillance to computer games and animation.

New product development (NPD) is vulnerable to a wide variety of risks arising from within the firm or from the external environment. Existing categorizations of NPD project risks are partial or ill-defined and consequently there is no clear consensus among researchers and practitioners about what constitute NPD project risks. To address this gap, this thesis deploys a systematic literature methodology to inductively develop a comprehensive risk taxonomy from a review of 124 empirical studies. This taxonomy is then empirically validated through a survey capturing data from 263 NPD projects conducted by UK firms. The thesis further investigated the moderating effect of NPD project type (incremental or radical), firm size (SMEs and large firms) and industry sectors on the proposed risk taxonomy. Variation in the perceptions of NPD risk by different members of the team was explored as well. The findings revealed that the principal risk factors affecting NPD projects are technological rapidity risk, supply chain risk, lack of funding and resource risk. The risk profile of radical NPD projects differed to that of incremental projects. SMEs were more vulnerable to NPD project risks than large firms. Most risks influenced NPD projects equally across industrial sectors. Members of NPD project teams from different backgrounds or with different roles perceived risks differently. The proposed taxonomy and its subsequent empirical validation provides a comprehensive and robust taxonomy for identifying and managing risks associated with different types of NPD project conducted by firms of varying sizes from different industrial sectors.

In the current era of state-of-the-art cutting edge technologies, businesses and organisations are rushing to transform their trade into e-business. The opportunity to utilise e-business improves their chances of gaining a larger market share by maximizing product availability, reducing the day-to -day business activity processing time, and providing related services in a convenient and inexpensive way to their customers. However in this race, the e-business growth pendulum is only swinging one way, and it is easy to understand the reason for this by observing today's business market. Due to the current financial condition, small business organizations (e.g. local retail shops) cannot afford costly IT systems and the associated maintenance/administration costs, but despite these financial constraints they have an overriding need for computing facilities in order to survive and compete with larger competitors by expanding their businesses. In this research, Web 2.0 and SOA (Service Oriented Architecture) technology are used to provide a middleware collaboration model between data persistence logic and an operation's requests. This layer helps to overcome the hard-coded service mapping with interface and generic customized workflow problems. This research further provides a mediation platform for request brokers and a high level of abstraction by encapsulating the low level details of the system. These are the most vital requirements to provide a platform, which have the capability of customizing business logic and handle both generic and customized workflows and subsequently to help SMEs (Small- to Medium-sized Enterprises) to convert their businesses to e-business swiftly at minimal cost.

The objective of this research is to develop a simulation software tool, GASFLO, which should evaluate pressure, flow and temperature distributions of process gas in pellet induration system networks. Pellet induration systems are complex industrial systems composed of heterogenous components. The magnitude of gas through leaks i.e. the air entering or leaving the system from the points other than the known exits, is substantial and it adversely effects the performance of induration process. These leaks are very difficult to measure because of the hostile environment in the plant. The modelling of such industrial systems requires a notable amount of experimentation so the tool has been designed to enable the user modeller to change the component models and solution algorithms easily. The conventional methods for flow network simulation are based on process centred approach, mostly composed of homogeneous components. For ease of computation, the non-pipe elements are modelled with an approximate linear or non-linear generic equation, whose coefficients can simulate different states of the element. The resulting set of non-linear equations is linearised and solved simultaneously using some iterative method. By contrast, GASFLO is based on device centred or unit based approach, and uses a two level hierarchical solution algorithm. The pellet induration system network is first idealised into a connected graph of streams (sets of serially connected components) and nodes. At the top or coordination level the flow and pressure distributions satisfying the Kirchhoff's laws are evaluated for the connected graph. At the lower or component level the exact mathematical models of components ale computed, in order of their occurrence in respective streams, using coordination variables as parameters. The converged flows are used for the temperature computation. The solution algorithm requires partitioning of the connected graph into forest and coforest structures, for which secondary algorithms have been developed using specific heuristics relevant to the pellet induration systems. The rigorous application of software engineering techniques for the design and implementation of software, enabled the resolution of the complexity of the modelled system, embedded the characteristics of 'quality software' into the resulting code and benefits from object orientation, even though it is implemented in standard FORTRAN 77. GASFLO predicted results are in a good agreement with the measured results, it has been validated for a real life pellet induration system. It has been applied to simulate several practical scenarios, like addition of extra wind boxes to the zones and to determine how the plant production can be increased by certain ratio, such simulations were not feasible otherwise. GASFLO takes less than a minute to simulate a real-life pellet induration system on a 486 PC. The combined simulation with an other software tool, INDSYS, which evaluates the heat distribution in the solids, is also feasible.

This thesis contains an account of the pwn(^+) final state produced by the interaction of 11.7 GeV/c π(^+) mesons in a hydrogen bubble chamber. 1 event for 4-prong DST = 0.17 µb. The total cross sections for 4-prong 1Cπº events and for the pwπ(^+) final state respectively are 1.31 ± 0.13 mb and 0.134 ± 0.O13 mb. These cross sections vary as a power of incident momentum with exponents of 0.82 and 1.41 respectively. For the reaction π(^+)p → wΛ(^++), values of mass and width of the Δ(^++) and total cross-section σ(π(^+)p → wΔ(^++)) respectively are 1234 ± 8 MeV/c(^2), 122 ± 32 MeV/c(^2) and 53 ± 10 µb. The cross-section decreases with the incident momentum as p(^-n)(_inc) with n = 1,85. The dσ/d|t| and dσ/d|t'| distributions show dips at |t| (or |t'|) = O and = 0.6 (GeV/c)(^2). The slope of the differential cross sections is = 4 GeV(^-2). The predictions on the density matrix elements disagree with the experimental results. There are dips at |t| = O and = 0.5 (GeV/c)(^2) for the natural parity contribution and a dip appears at |t| = 0.2 (GeV/c)(^2) for the unnatural component. Most of the values of the double statistical tensors are compatible with zero. A qualitative theoretical discussion of the results is presented. For the reaction π(^+)p → pB(^+), values of mass, width and total cross-section of the B(+) respectively are 1235 ± 12 MeV/c(^2), 120 ± 48 MeV/c(^2) and 21 ± 6 µb. The cross-section decreases with the incident momentum as p(^-n)(_inc) with n about 1.5. The slope of the differential cross-section is = 4 GeV(^-2). The helicity amplitude for the wº, |f(_o)|(^2) = 0.09 ± 0.07 which disagrees with the prediction of a simple quark model. The values of favour J(^P) =1(^+) or 2(^+) but disfavour J = 3 for the B. Assuming J(^P) = 1(^+) assignment, the density matrix elements show that the B(^+) is weakly aligned both in the t- and the s-channel helicity frames. There is no evidence for the production of the higher mass Δ(^++) or higher mesons decaying into wπ(^+) or any resonance decaying into wp.

By thoroughly analyzing 910 U.S. House immigration bills from the 113th, 114th, 115th, and 116th Congressional sessions, my Ph.D. research delves into the impact of health crises on introduced U.S. House immigration bills in the United States. My research fills a crucial gap in the literature by examining the influence of legislative policy entrepreneurs (LPEs) on agenda-setting and socio-political discourse in health crises. By using Kingdon's policy entrepreneur theory and the inductive qualitative method of relational content analysis, I explore the general theme, underlying tone, rhetoric, and proposed measures of House immigration bills during health crisis versus non-health crisis periods. The findings reveal that elected House representatives are more likely to introduce restrictive immigration bills during health crises and that geographical location and political affiliation play a significant role in shaping these bills' rhetoric and proposed measures. Using the cumulative ANES dataset from 1948 to 2020, I demonstrate that the general population tends to be less welcoming towards immigrants and favors more restrictive immigration policies during health crises. Political ideology, education, income scale, and gender significantly determine public attitudes toward immigration policies and immigrants. My research sheds light on economic conditions, political environment, and legal frameworks that influence the legislative activity of elected House representatives and the public's attitudes toward immigration policy. The findings provide valuable insights and directions for future research, policy, and practice efforts toward a more equitable and just society.

Abstract: An alarm system is an integral part of process monitoring and safety. A poorly configured alarm system can sometimes cause more harm than good, by introducing many false and nuisance alarms for the operators. Various standards on alarm system management and rationalization suggest many configuration methods to help in improving the overall performance of alarm monitoring systems. In this thesis, the problem of analyzing and designing alarm systems for both single- and multi-mode processes is considered. A design procedure of a multivariate alarm system for multi-mode processes is developed. A hidden Markov model based modeling approach is adopted to capture the multi-modality of data and the mode-reachability constraints of a multi-mode process. A monitoring index utilizing the proposed two-step Viterbi algorithm is developed, and for fault isolation, reconstruction based contribution plots are used. The utility of delay-timers in improving existing univariate alarm systems for multi-mode processes is studied. A mathematical model is developed to calculate analytical expressions for different performance indices (the false alarm rate, missed alarm rate, and expected detection delay). A particle swarm optimization based method is proposed for designing delay-timers, while satisfying the constraints on the performance indices and delay-timer lengths for various modes of the operation of a process. The analysis and design of time-deadbands for univariate alarm systems is also considered in this thesis. In particular, a Markov chain process based mathematical model is developed to capture the time-deadband configurations for single mode processes. Analytical expressions for the performance indices are calculated, and design procedures based on process data and alarm data are developed.

Complex processes of perception, cognition and action are thought to emerge from the self-organized dynamics of different brain networks. Humans and many different animals exhibit a vast repertoire of complex movements in response to environmental stimuli. The neural basis of this voluntary motor control in particular is not very well understood. Understanding the basis of these movements is of interest to biologists, neuroscientists, engineers and computer scientists. Once modeled, they can be useful in applications such as robotics, biomorphic chips, and machine learning. Spatiotemporal patterns of neural activity have increasingly come to be seen as important for encoding information in the nervous system – especially with respect to the control of movement. For a given individual, any specific voluntary movement is ultimately encoded as an aperiodic spatiotemporal pattern of activation across a set of muscles, and presumably in spinal, and perhaps cortical, motor neurons. Different individuals can be seen as having different stereotypical patterns of activity which determine the “style” of specific movements – e.g., how they walk or write. Experimental studies also indicate that these activity patterns may themselves be constructed as linear combinations of a few fixed spatiotemporal basis patterns of activity called motor synergies. For this to work, it is essential that neural systems be able to represent spatiotemporal activity patterns that are stimulus-specific, aperiodic (i.e., not rhythmic), transient (i.e., lasting only briefly), and robust (i.e., at least somewhat tolerant of errors and noise). In this thesis, a simple recurrent neural network model is introduced motivated by the need to understand the basis of voluntary motor control, arising as a result of synergies. Distinct attractors representing spatiotemporal patterns of activity are embedded in the system having efficient storage and recall. Since the model is motivated by a need to understand motor control, the patterns are also visualized by turning them into observable movement trajectories in space. Different attractors elicit distinct trajectories. The attractors can also be triggered in combination to give new and complex spatiotemporal patterns, resulting in novel trajectories representing movements. The system is shown to exhibit many of the properties needed for the flexible construction of complex, aperiodic movements.