Fault detection and diagnosis of gear transmission systems have attracted considerable attention in recent years, due to the need to decrease the downtime on production machinery and to reduce the extent of the secondary damage caused by failures. However, little research has been done to develop gear shaft and planetary gear crack detection methods based on vibration signal analysis. In this thesis, an approach to gear shaft and planetary gear fault detection based on the application of the wavelet transform to both the time synchronously averaged (TSA) signal and residual signal is presented. Wavelet approaches themselves are sometimes inefficient for picking up the fault signal characteristic under the presence of strong noise. In this thesis, the autocovariance of maximal energy wavelet coefficients is first proposed to evaluate the gear shaft and planetary gear fault advancement quantitatively. For a comparison, the advantages and disadvantages of some approaches such as using variance, kurtosis, the application of the Kolmogorov-Smirnov test (K-S test), root mean square (RMS) , and crest factor as fault indicators with continuous wavelet transform (CWT) and discrete wavelet transform (DWT) for residual signal, are discussed. It is demonstrated using real vibration data that the early faults in gear shafts and planetary gear can be detected and identified successfully using wavelet transforms combined with the approaches mentioned above. In the second part of the thesis, the planetary gear deterioration process from the new condition to failure is modeled as a continuous time homogeneous Markov process with three states: good, warning, and breakdown. The observation process is represented by two characteristics: variance and RMS based on the analysis of autocovariance of DWT applied to the TSA signal obtained from planetary gear vibration data. The hidden Markov model parameters are estimated by maximizing the pseudo likelihood function using the EM iterative algorithm. Then, a multivariate Bayesian control chart is applied for fault detection. It can be seen from the numerical results that the Bayesian chart performs better than the traditional Chi-square chart.

Three stages of research were carried out to investigate the use of the self-paced visual occlusion technique, and to model visual information sampling. Stage 1. A low-fidelity driving simulator study was carried out to investigate the effect of glance duration, a key parameter of the self-paced occlusion technique, on occlusion times. Results from this experiment, paired with analysis of data available from an on-road driving study, found an asymptotic relationship between the two variables. This finding has practical implications for establishing the appropriate glance duration in experimental studies that use self-paced visual occlusion. Stage 2. A model of visual information sampling was proposed, which incorporates elements of uncertainty development, subjective thresholds, and an awareness of past and current states of the system during occlusion. Using this modelling framework, average information sampling behaviour in occlusion studies can be analysed via mean occlusion times, and moment-by-moment responses to system output can be analysed via individual occlusion times. Analysis using the on-road driving data found that experienced drivers demonstrated a more complex and dynamic sampling strategy than inexperienced drivers. Stage 3. Findings from Stage 2 led to a simple monitoring experiment that investigated whether human operators are in fact capable of predicting system output when temporarily occluded. The platform was designed such that the dynamics of the system naturally facilitated predictions without making the monitoring task trivial. Results showed that participants were able to take predictive information into account in their sampling decisions, in addition to using the content of the information they observed from each visual sample.

Three stages of research were carried out to investigate the use of the self-paced visual occlusion technique, and to model visual information sampling. Stage 1. A low-fidelity driving simulator study was carried out to investigate the effect of glance duration, a key parameter of the self-paced occlusion technique, on occlusion times. Results from this experiment, paired with analysis of data available from an on-road driving study, found an asymptotic relationship between the two variables. This finding has practical implications for establishing the appropriate glance duration in experimental studies that use self-paced visual occlusion. Stage 2. A model of visual information sampling was proposed, which incorporates elements of uncertainty development, subjective thresholds, and an awareness of past and current states of the system during occlusion. Using this modelling framework, average information sampling behaviour in occlusion studies can be analysed via mean occlusion times, and moment-by-moment responses to system output can be analysed via individual occlusion times. Analysis using the on-road driving data found that experienced drivers demonstrated a more complex and dynamic sampling strategy than inexperienced drivers. Stage 3. Findings from Stage 2 led to a simple monitoring experiment that investigated whether human operators are in fact capable of predicting system output when temporarily occluded. The platform was designed such that the dynamics of the system naturally facilitated predictions without making the monitoring task trivial. Results showed that participants were able to take predictive information into account in their sampling decisions, in addition to using the content of the information they observed from each visual sample.

In this study, a user-centred design approach was used to develop a mobile health application designed to support adult T1DM patients with their self-management routine. In the requirements gathering phase, an observational study of a diabetes clinic and patient interviews were conducted. An analysis of the data collected from this phase helped identify the functional design requirements used to guide the design. Using a rapid prototyping approach, data visualizations, game-based elements, carb-counting and social networking features were explored. The final prototype developed in this research was evaluated for its ease of use and perceived usefulness. The design was found to be generally easy to use. With respect to data visualizations, participants preferred the scatter plot view of their blood glucose readings to a bar chart. In addition, it was found that all participants wanted a way to track their HbA1c on a regular basis.

Creativity is important in the design and manufacture of successful products, yet neither creativity nor the early stages of design are well understood. This lack of understanding limits the tools that can be developed to support the crucial earlier stages of design that ultimately determine product success. My research aims to better understand creativity by studying and quantifying the potential of semantic stimuli (words) presented during concept generation. Natural language was chosen as design stimuli because language provides a systematic framework for stimuli generation. Furthermore, natural language is ubiquitous and intimately related to cognitive functions required in design such as reasoning and memory. Ultimately, the results of this research will assist in the development of early-design support tools. In a series of four experiments, the effects of semantic stimuli oppositely and similarly related to the experiment problem were examined with respect to creativity and designers’ language patterns. Results show that opposite-stimulus concepts were significantly more creative than similar-stimulus concepts. It also was observed that opposite stimuli elicited designer behaviours that may encourage creative concepts. These results suggest that the use of oppositely related stimulus words is a practical method for encouraging creative design.

Creativity is important in the design and manufacture of successful products, yet neither creativity nor the early stages of design are well understood. This lack of understanding limits the tools that can be developed to support the crucial earlier stages of design that ultimately determine product success. My research aims to better understand creativity by studying and quantifying the potential of semantic stimuli (words) presented during concept generation. Natural language was chosen as design stimuli because language provides a systematic framework for stimuli generation. Furthermore, natural language is ubiquitous and intimately related to cognitive functions required in design such as reasoning and memory. Ultimately, the results of this research will assist in the development of early-design support tools. In a series of four experiments, the effects of semantic stimuli oppositely and similarly related to the experiment problem were examined with respect to creativity and designers’ language patterns. Results show that opposite-stimulus concepts were significantly more creative than similar-stimulus concepts. It also was observed that opposite stimuli elicited designer behaviours that may encourage creative concepts. These results suggest that the use of oppositely related stimulus words is a practical method for encouraging creative design.

In an emergency department (ED), information flow is of high value, as the ability to react quickly directly affects the patients’ well being. One of the gaps in the information flow is in the order tracking process. This paper focuses on modelling the feedback in this process from the order being issued until it has been fulfilled. We address this problem using discrete-event simulation. Additionally, we use the mathematical theory of communication to evaluate the information con- tent in the current and proposed systems. We perform computational tests on these models to compare their performance. Experimental results show that the problem can be effectively modelled using our approach and the effects of feedback on the physician decision-making can be better understood. The results indicate that additions of as little as one point of feedback have practically significant effects on the amount of time that an order spends in the system.

We explore both robust biologically guided intensity-modulated radiation therapy (BG-IMRT) and pattern recognition to identify responders to cancer treatment for lung cancer. Heterogeneous dose prescriptions that are derived from biological images are subject to uncertainty, due to potential noise in the image. We develop a robust optimization model to design BG-IMRT plans that are de-sensitized to uncertainty. Computational results show improvements in tumor control probability and deviation from prescription dose compared to a non-robust model, while maintaining tissue dose below toxicity levels. We applied machine learning algorithms to 4D gated positron emission tomography/computed tomography (PET/CT) scans. We identified classifiers which could outperform a naive classifier. Our work shows the potential of using machine learning algorithms to predict patient response. This could hopefully lead to more adaptive treatment plans, where the clinician would adapt the treatment based on the prediction provided at certain time intervals in the treatment.

We explore both robust biologically guided intensity-modulated radiation therapy (BG-IMRT) and pattern recognition to identify responders to cancer treatment for lung cancer. Heterogeneous dose prescriptions that are derived from biological images are subject to uncertainty, due to potential noise in the image. We develop a robust optimization model to design BG-IMRT plans that are de-sensitized to uncertainty. Computational results show improvements in tumor control probability and deviation from prescription dose compared to a non-robust model, while maintaining tissue dose below toxicity levels. We applied machine learning algorithms to 4D gated positron emission tomography/computed tomography (PET/CT) scans. We identified classifiers which could outperform a naive classifier. Our work shows the potential of using machine learning algorithms to predict patient response. This could hopefully lead to more adaptive treatment plans, where the clinician would adapt the treatment based on the prediction provided at certain time intervals in the treatment.

Adaptive radiation therapy re-optimizes treatment plans based on updated tumor geometries from magnetic resonance imaging scans. However, the imaging process is costly in labor and equipment. In this study, we develop a mathematical model that describes tumor evolution based on a Markov assumption. We then extend the model to predict tumor evolution with any level of information from a new patient: weekly MRI scans are used to estimate transition probabilities when available, otherwise historical MRI scans are used. In the latter case, patients in the historical data are clustered into two groups, and the model relates the new patient's behavior to the existing two groups. The models are evaluated with 33 cervical cancer patients from Princess Margaret Cancer Centre. The result indicates that our models outperform the constant volume model, which replicates the current clinical practice.