Filters
Availability
Publication Year
Resource Type
Faculty
School/Institute
Publication Search Results
Sort
Results per page
1 - 10 of 108
-
(2021)ThesisNavigation is a technique for the determination of position and attitude of a moving platform with respect to a known reference. Global Navigation Satellite System (GNSS) has become a dominating navigation technology. However, GNSS signals are degraded or denied in indoor environments. It is necessary to develop alternative positioning techniques for indoor navigation to realize seamless navigation. Inertial Navigation System (INS) and Vision are both regarded to be highly promising because of their ubiquitous and self-contained nature. A new indoor navigation system with vision, INS and reality-based 3D maps are proposed. The main contributions of this thesis are summarized as follows: 1. A new strategy for the integration of vision with a low-cost INS has been developed based on a smartphone. This new approach solves the difficulty of precise calibration of INS errors in such a scenario and enables MEMS INS to generate stable position and attitude solutions. 2. Results show that improved accuracy and reliability of the geo-referenced solution can be achieved. Vision-based navigation with reality-based 3D maps (Vision)/INS integration improves the accuracy and robustness of a navigation solution compared with an INS only solution. 3. Aiding Optical Flow (OF) and Visual Odometry (VO) navigation solution to Vision/INS integrated system improved geo-referenced results during Vision outages. The results confirmed the effectiveness of integration for high accuracy positioning applications. 4. A novel geo-referenced system based on Vision/OF/VO/INS integration has been developed and tested for indoor navigation. Real experiments are conducted to evaluate the influence of different integrated configurations on the performance of the navigation system. 5. Integrated indoor navigation requires a robust outlier detection mechanism to ensure good performance. Outlier detection and identification are explored and researched on the integrated indoor navigation system. Besides, a multi-level outlier detection scheme for the navigation system has been proposed. 6. Analyzing the factors that influence the correlation coefficients between fault test statistics in Vision/INS measurements and the dynamic model is another essential contribution of this thesis. Reliability and separability analysis of outlier detection theory was extended by providing a more reliable estimation of MDB and MSB.
-
(2021)ThesisDue to the large variety and unique physiochemical properties, such as high electrical conductivity, adjustable interlayer spacing, intercalation chemistry and so on, two-dimensional (2D) materials have attracted tremendous attention for their suitability in the development of high-performing supercapacitors. Despite that great progress has been made, there is still no single 2D material that can perfectly meet all the requirements to replace the existing material, mainly activated carbon, used in commercial supercapacitors. Therefore, continuing efforts for exploring novel, high-performing 2D materials with low cost are desirable. In this dissertation, the prior studies on the development of 2D materials ranging from layered inorganic materials to organic-inorganic hybrids for supercapacitors are first reviewed. As an emerging type of 2D material, layered organic-inorganic hybrids start to show promising results to be used to fabricate high-density, nonporous, and thick electrodes for compact capacitive energy storage. However, the studies in this area are still lacking. Thus, the goal here is to explore the opportunities of 2D organic-inorganic hybrids for applications in supercapacitors. The relevant techniques and methods used throughout the study are then outlined. Next, three research chapters supporting the main findings of the investigation are included. The first research chapter describes a facile mechanical strategy to improve the kinetics and rate performance of 2D organic-inorganic hybrid electrodes at ultrahigh mass loadings (up to 30 mg cm-2). The second research chapter reports the synthesis of a new layered organic-inorganic hybrid material with excellent volumetric and areal capacitances even at mass loadings reaching 50 mg cm-2, highlighting the good electrode kinetics. The third research chapter presents the wafer-scale electrochemical deposition synthesis of 2D organic-inorganic composite films with controlled size and thickness, which are promising for the development of flexible and transparent electrochemical energy storage devices. Finally, conclusions and recommendations are given at the end of this dissertation.
-
(2021)ThesisUnderwater sound can have a detrimental effect on marine animals due to the ever-increasing noise levels in their pristine habitat. It has also been commonly used to detect underwater floating objects via a sonar system. To absorb unwanted underwater sound, polymers (e.g., rubber), which have similar impedance to that of water, are widely used for sound absorption in water. Nanocomposites have attracted considerable attention due to their ability to improve sound absorption properties of polymer-based sound absorption materials. This project aims to develop a thin-layer nanocomposite with high underwater sound absorption at low frequency and high pressure. A water-filled impedance tube, an essential facility to test new materials developed in this PhD thesis, was designed and constructed. The established research facility consists of four main components: a stainless steel tube and its supporting devices, a sound source (a projector) and its associated electronics, an underwater sound pressure measurement system, and a water pressurized system. Subsequent calibrations and measurements showed that the established apparatus could be used to measure the underwater sound absorption coefficient in a frequency range of 1500 Hz to 7000 Hz and under hydrostatic pressure in a range of 0 to 1.5 MPa. Carbon nanotubes (CNTs) reinforced polydimethylsiloxane (PDMS) nanocomposites were designed, fabricated, and tested. This development comprised of two stages. In the first stage, PDMS was selected as the material matrix, surfactant and carboxyl functionalized multi-walled carbon nanotubes (MWCNT-COOH) as inclusions, and a new nanocomposite, namely PSM (PDMS/surfactant/MWCNT-COOH), was then developed. Effects of the added surfactant and MWCNT-COOH on the mechanical properties, chemical properties, and morphology were investigated, which indicated the nanocomposite’s potential for sound absorption improvement. Underwater acoustic tests showed high underwater sound absorption coefficients (>0.8) in the most frequency range 1500 Hz to 7000 Hz. However, it was observed that a significant drop in the underwater sound absorption performance under high hydrostatic pressure. It was found that the high compression of PSM was the cause of poor performance under high hydrostatic pressure. In the second stage, a core-shell structure was designed to maintain the high sound absorption coefficient of PSM under high hydrostatic pressure. A novel structure of a 2-mm-thick hard shell with a 2-mm-thick soft layer was developed to encapsulate the PSM sample so that its deformation can be minimized and its superior sound absorption property was improved under high pressure. Experimental results on the water-filled impedance tube demonstrated that the new structure offered a promising solution to the demand for advanced underwater materials, which are thin and have high sound absorption performance under high hydrostatic pressures. In summary, this study has developed a polymer-based nanocomposite. Mechanical properties, chemical properties, morphology, and underwater acoustic properties of the nanocomposite have been studied. The nanocomposite is thinner than existing underwater acoustic materials and has excellent underwater sound absorption performance in the frequency range of 1.5 to 7 kHz and under atmospheric pressure. For applications in high hydrostatic pressure up to 1.5 MPa, the proposed new structure with a total thickness of 14 mm, in comparison to 50 mm or more thickness of other developed materials for marine applications, showed good sound absorption results and potentially addressing the on-going technical challenge of poor sound absorption performance of acoustic materials under high hydrostatic pressure.
-
(2021)ThesisCataracts are responsible for almost half of worldwide blindness, making it one of the biggest health challenges in this era. Cataracts are irreversible because of their pathology, which is controlled by the aging and biochemical change of eye tissues. As a result cataract surgery is currently the only effective treatment. The general procedure of cataract surgery includes separation and removal of the failed lens tissue from the surrounding soft tissue in the eye, followed by artificial lens implantation. Lens removal requires successful separation of lens tissues as a critical step that determines surgical success. However key parts of cataract separation affected by fluid mechanics and rheology are uncharacterised. This project aims to explain the behaviors of such separation phenomena and connect fundamentals with possible explanations and enhancements. A multi-layer bio-polymer injection model is developed to mimic the separation process in cataract surgeries. The separation can be considered peeling of a soft elastic tissue by a pressure-driven fluid flow, whose performance is closely related to properties such as flow rate and velocity as well as fluid viscosity, normal stress and yield stress. In our project, the separation physics is studied as a hydraulic fracture problem. Theories are proposed to discuss the effectiveness and safety of hydraulic fracture with different flow and fluid parameters. It is found both higher flow rate and viscosity will cause tissue to be deformed more, which may increase the risks of tissue damage. Yield stress fluids with significant elasticity are not suitable as in most cases they rupture the tissue. Normal stress fluids have the potential to provide safe and effective separation. It is found that with a small scale separation, however, the separation effectiveness is mainly affected by the flow rate, and the fluid properties play a more minor role. General ideas and potential improvements according to our results and theories are also proposed for cataract surgeries, which we hope will contribute to easier and safer separation.
-
(2021)ThesisThe complex structure of software applications can increase cognitive load and render tools incomprehensible. Since few studies have been conducted that focus on facilitating the learning of software applications for novice users, this thesis proposed a teaching solution by applying three elements of gamification including the use of Narrative, Interactivity and Avatar. The goal was to apply these gamification elements in an e-learning system and evaluate the effects on learners’ cognitive load while learning to use software tools with low and high element interactivity. Cognitive load theory was used as a guiding research principle. To this end, three integrated experiments were designed with the total of 160 participants. A mixture of objective and subjective quantitative measurement methods was used to measure cognitive load. For the subjective measurement, participants were asked to complete a self-reported difficulty Likert scale questionnaire. For the objective measure, participants performance including the following five factors was assessed: test task performance marks; test task performance speed; mouse movement distance; number of left and right clicks while finding the test task solution; time duration of reading each tutorial. In the first experiment, narrative which is a core element of gamification science, was selected as a procedure that can provide practical knowledge to software learners while impacting cognitive load by providing a familiar theme in worked-examples. The results showed that an e-learning system with a familiar narrative could decrease cognitive load in comparison to the no-narrative and unfamiliar narrative systems. In the second experiment, the effect of interactivity on delivering narrative-based content was evaluated by comparing animation versus interactive animation. The findings revealed that interactive animation was superior to the animation-based version which is in accord with embodied cognition theory. Finally, the third experiment evaluated the effect of a talking avatar versus plain audio on cognitive load in narrative-based e-learning systems that used interactive animation. The findings indicated that the talking avatar increased cognitive load during learning which is in accord with the redundancy effect.
-
(2021)ThesisIn the recent decade there has been a sharp increase in the utilization of machine/deep learning models for the development of Natural Language Processing (NLP) applications, especially focused on language understanding, with end-goals targeted at, but not limited to: information retrieval, machine translation, sentiment analysis, question answering, etc. These applications call for the need of better models for in-depth understanding of the language structures which in turn help development of automated routines that may acquire vast variety of unstructured data from web resources, process the data and convert it to the desired information content. In the recent past many different models have been developed for language-specific information extraction and a better understanding of semantic aspects of the language, however, yet there are some challenges associated that need to be addressed for the improved utility of these models in the down-streaming tasks. In this thesis, we propose new models with the aims to improve upon the existing methods for the lexico-semantic relation and information extraction tasks. For lexico-semantic relation extraction, we work around distinguishing among different lexico-semantic relations captured from unstructured data, i.e., distinguishing antonyms from synonyms and hypernymy detection. For information extraction, we work with improving the Fine-Grained Named Entity Typing (FG-NET), which is a key component for different down-streaming information and relation extraction tasks. Given the fact that the fine-grained type hierarchy follows a hierarchical structure, in Chapter 5, we extend the concepts of FGET-RR to Fine-Grained Named Entity Typing with Refinement in Hyperbolic space (FGNET-RH) that combine the benefits of the non-euclidean geometry (hyperbolic space) along with the graph structures to perform FG-NET in performance-enhanced fashion. Finally, in order to evaluate the reliability of the machine/deep learning systems for information extraction in real-life scenarios, we evaluate the performance of these systems under uncustomized settings.
-
(2021)ThesisStability of power system ensures continuous and uninterrupted power delivery from generation to demands and measures the dynamic behaviour of different system characteristics subject to disturbances. Failure to maintain stability can lead to consequences ranging from disconnected buses and areas to cascading failures and system blackouts. Increased integration of renewable energy sources and involvement of demand-side into system operation in the recent years have introduced more intermittency into the power system. This intermittency increases the difficulty to maintain system stability and requires better control strategies to accommodate the high-level penetration of uncertainties in the power system operation. In addition to conventional control methods such as generator dispatch, numerous measures are considered to reduce the uncertainty introduced, such as wind curtailment, to act timely in case of instability, such as emergency load shedding, and also to utilize the flexible resources, such as energy storage systems. This research consists of methodologies that account for control measures both before and after the contingency to achieve the optimal balance between the economy and stability in system operation. The main research contributions are listed below: Firstly, a hybrid approach is proposed for the improvement of conventional optimal power flow (OPF) calculation, which utilizes the critical machine (CM) theory in the Extended Equal Area Criteria (EEAC). A detailed composite load model is later applied to capture a more accurate dynamic behaviour of the system. Secondly, a preventive-emergency coordinated control method is developed to maintain system transient stability in the presence of uncertain wind power generation. It coordinates between preventive control (generator dispatch and wind curtailment) and emergency control (emergency demand response) to achieve the optimal trade-off between system stability and economy. Thirdly, the effects of having actively controlled utility-scale Battery Energy Storage Systems (BESSs) systems are investigated, and a coordinated strategy that utilizes generator redispatch in preventive control and energy storage in both preventive and emergency control stage are developed. Finally, the flexibility of BESS is investigated. The model performance and parameter sensitivity of a new generic BESS model are firstly evaluated. An emergency control scheme is designed based on the results of evaluation, which includes both BESS and emergency load shedding scheme as post-contingency control measures. Proposed methodologies and strategies are evaluated on benchmark systems with industry-grade dynamic models and simulation software. Applicable cases from literature are also performed for comparative purpose. The simulation experimental results have verified the effectiveness of the proposed methodologies and strategies.
-
(2021)ThesisA bipartite graph is a graph with two layers such that vertices in the same layer are not connected, which is widely used to model the relationships among two types of entities. Examples of bipartite graphs include author-paper networks, customer-product networks, and ecological networks (e.g., the predator-prey network and the plant-animal network). In bipartite graphs, cohesive subgraph computation is a fundamental problem that aims to find closely-connected subgraphs, which can be applied to group recommendations, network visualization, and fraud detection. In this thesis, we propose a novel cohesive subgraph model called τ -strengthened (α, β)- core (denoted as (α, β)τ -core), which is the first work to consider both tie strength and vertex engagement on bipartite graphs. An edge is a strong tie if contained in at least τ butterflies (2 x 2-bicliques). (α, β)τ -core requires each vertex on the upper or lower level to have at least α or β strong ties, given strength level τ. To retrieve the vertices of (α, β)τ -core optimally, we construct index Iα,β,τ to store all (α, β)τ -cores. Effective optimization techniques are proposed to improve index construction. To make our idea practical on large graphs, we propose 2D-indexes Iα,β, Iβ,τ, and Iα,τ that selectively store the vertices of (α, β)τ -core for some α, β, and τ . The 2D-indexes are more space-efficient and require less construction time, each of which can support (α, β)τ -core queries. As query efficiency depends on input parameters and the choice of 2D-index, we propose a learning-based hybrid computation paradigm by training a feed-forward neural network to predict the optimal choice of 2D-index that minimizes the query time. Extensive experiments show that (1) (α, β)τ -core is an effective model capturing unique and important the proposed techniques significantly improve the efficiency of index construction and query processing.
-
(2021)ThesisRelation prediction is a fundamental task in network analysis which aims to predict the relationship between two nodes. Thus, this differes from the traditional link prediction problem predicting whether a link exists between a pair of nodes, which can be viewed as a binary classification task. However, in the heterogeneous information network (HIN) which contains multiple types of nodes and multiple relations between nodes, the relation prediction task is more challenging. In addition, the HIN might have missing relation types on some edges and missing node types on some nodes, which makes the problem even harder. In this work, we propose RPGNN, a novel relation prediction model based on the graph neural network (GNN) and multi-task learning to solve this problem. Existing GNN models for HIN representation learning usually focus on the node classification/clustering task. They require the type information of all edges and nodes and always learn a weight matrix for each type, thus requiring a large number of learning parameters on HINs with rich schema. In contrast, our model directly encodes and learns relations in HINs and avoids the requirement of type information during message passing in GNN. Hence, our model is more robust to the missing types for the relation prediction task on HINs. The experiments on real HINs show that our model can consistently achieve better performance than several state-of-the-art HIN representation learning methods.
-
(2021)ThesisThis thesis presents an experimental investigation of air, argon and helium sonic under-expanded jets transversely injected through a turbulent boundary layer into a supersonic Mach 3 flow. The aims of the thesis are to demonstrate the steady and unsteady flow features of Transverse Jets In Supersonic Crossflow (TJISC), to present the shear layer vortex shedding frequency and the related controlling parameters, to extract the dominating coherent structures, and to present the mean and fluctuating pressure loads on the wall around the jet port beneath the TJISC. The flow field was visualized using two types of schlieren methods, meanwhile, the mean and pressure fluctuation distributions beneath the TJISC were measured by Pressure-Sensitive Paint (PSP) and a high-speed pressure transducer array, respectively. Besides the general flow features, instantaneous schlieren images reveal the unsteady nature of the TJISC. The quasi-periodically shedding shear layer vortices interact with the adjacent shock system and cause intense quasi-periodical deformation of the shock system. The Mach disk and the barrel shock presented in the air and argon cases are absent in the helium cases. In the convective frame for the helium cases, these shear layer vortices travel at supersonic speed and generate a series of moving shock waves that are propagating along the shear layer. The penetration depth of the helium TJISC is slightly higher than the air and argon cases due to these moving shocks. Power Spectral Density (PSD) of schlieren image pixel light intensity shows that the peak frequency of vortex shedding is inversely proportional to the momentum flux ratio J and this may be due to the level of compressibility. At the same J, the peak vortex shedding frequencies of the air and argon TJISC are similar, while the frequency of the helium TJISC is approximately double. Spectral Proper Orthogonal Decomposition (SPOD) and Dynamic Mode Decomposition (DMD) were applied to the schlieren data and coherent structures were extracted. The SPOD results show that the modal energy peak frequencies are consistent with the shear layer vortex shedding frequency, and the first mode that represents the shear layer vortices contains most of the modal energy. The SPOD results indicate that the flow field is relatively low-rank, and the shed vortices in the shear layer are dominant. Pressure fluctuations along the centre line beneath the jet illustrate that signals of the most upstream transducer (upstream of jet port) are dominated by the separated boundary layer. The signals of the second upstream transducer are dominated by the fluctuations of the shear layer vortices and shock structures in the air and argon cases, while the signals of the helium cases are relatively broadband. At the most downstream locations, the PSD of the pressure fluctuations presents peaks that are generated by the wake structures near the wall. Pressure-sensitive paint results identify the high-pressure regions upstream of the jet port that are caused by the separation shock and the bow shock. A symmetric low-pressure region, a collision shock, and the wake structures are observed downstream of the jet port. In conclusion, the TJISC is an unsteady flow field with complex fluid mechanics that are closely linked to the injectant gas properties. The peak shear layer vortex shedding frequency is inversely proportional to the J. This conclusion is confirmed by the SPOD and DMD data and the extract modes are addressed. Pressure loads beneath the jets were presented and linked to the unsteady flow structures and injected gas properties. This thesis provides detailed information on the TJISC and can provide some insights on designing scramjet engines.