Carleton University Institutional Repository
Not a member yet
20816 research outputs found
Sort by
Partisans, Plutonium, and the Polygon: The Domestic and Global Impact of the Nevada-Semipalatinsk Movement
No full textThis thesis is about the Nevada-Semipalatinsk Movement (NSM), an anti-nuclear environmental movement that formed in Kazakhstan at the height of perestroika in 1989. It argues that the NSM had a notable impact on local and global politics. This thesis demonstrates that the NSM tapped into concerns about pollution that had existed among the Kazakh intelligentsia and ordinary citizens since the Brezhnev years. It shows how the NSM emerged in a period of heightened political and environmental tension. This thesis highlights how the organization was driven by nationalistic ideas and concerns about health and environmental sustainability. It traces the organization’s activities in Kazakhstan into the post-Soviet period. On a global scale, this thesis discusses the movement’s links to “citizen diplomacy” at the end of the Cold War. Analyzing the relationship between American and Soviet-Kazakh activists, this thesis reveals that citizen diplomacy was a complex and long-lasting phenomenon
Instructor Perceptions of e-Proctoring Software
No full textThe COVID-19 pandemic motivated higher education institutions to adopt the use of e-proctoring software as a means to maintain academic integrity. This study explores the tension between student privacy and academic integrity from instructors’ perspective. Through semi-structured qualitative interviews with 19 university instructors, our findings delineate the competing factors influencing instructors’ adoption or avoidance of e-proctored assessments: academic integrity, the online remote format, logistical considerations such as class size, departmental policies, and privacy considerations. We analysed instructors’ specific privacy attitudes towards e-proctoring, and perspectives regarding student privacy and institutional data protection practices. Lastly, we evaluated instructors’ appraisals on the efficacy of e-proctoring software. We found that most instructors deprioritized privacy considerations when deciding whether to adopt e-proctoring, but viewed academic integrity as the topmost inviolable priority. Overall, we provide insight into the complexities of managing privacy and competing priorities from the instructors’ perspective, and offer recommendations for instructors and institutions
Efficient Computations of Interesting Paths
No full textGiven a high dimensional dataset and a function of interest, the Mapper algorithm outputs a topologically accurate summary. This summary helps identify subpopulations, which motivates the use of this algorithm for data mining. Continuing the work on interesting paths by Kalyanaraman, Kamruzzaman and Krishnamoorthy, we study three optimization problems to maximize the total interestingness score of the paths on the Mapper output. The Max-IP problem is solved in directed acyclic graphs, but we extend the solution to a special class of graphs which is common for the Mapper algorithm. For the k-IP problem, we show its preceding NP-completeness proof has gaps. We give a new NP-completeness proof of the k-IP problem for k ≥ 4. We design approximation and exact algorithms for it. For the IP problem, we give a proof of its NP-completeness. We also design corresponding approximate algorithms using various heuristics
Rapid Workflow for Energy Model Calibration, Retrofit Optimization, and Uncertainty Analysis of Large Building Energy Retrofits
No full textThe building sector has been targeted to have net zero emissions by 2050. Considering the low rate of new construction and degradation of the building's materials properties over time, the retrofit of existing buildings can play a crucial role in achieving these goals. Moreover, determining a cost-effective strategy for reducing GHG emissions is necessary. Due to budget and time limitations, determining the optimal retrofit and operating strategy to minimize GHG emissions, energy usage, and the life-cycle cost is vital for building decarbonizing. This research aims to develop a rapid framework for creating a retrofit roadmap using multi-objective optimization of building energy retrofits. The proposed thesis is divided into four main parts to achieve these objectives. First, an optimization algorithm calibrates the energy model using various measured data resolutions. The metered data of a large office building in Ottawa, Canada, are used as a case study. The second part develops a rapid framework for multi-objective optimization of building retrofit considering life cycle cost and GHG emissions. The proposed workflow is applied to an office building's retrofit analysis in Ottawa, Canada. Then, a methodology is developed to create a roadmap for building energy retrofitting based on the uncertainty analysis results. In this step, the uncertainty of the optimal strategy due to the fluctuation of some economic rates is evaluated using the Monte Carlo simulation. Finally, a hybrid model is proposed for retrofit optimization. This approach integrates physical modelling (EnergyPlus) with advanced data-driven approaches (machine learning) to optimize retrofit strategies. Utilizing machine learning techniques, particularly the Random Forest model, significantly improves accurate predictions and efficient retrofit optimization. This methodology, grounded in a comprehensive framework, showcased its potential to expedite decision-making while eliminating reliance on building simulation tools
Facilitating Programming-Based 3D Computer-Aided Design Using Bidirectional Programming
No full text3D Computer-Aided Design (CAD) applications allow users to create visual representations of models, helping create, edit, test, and analyze designs. Most offer a Graphical User Interface (GUI) with direct manipulation, providing easy-to-use interactions, while a less popular category adopts a programming-based approach requiring users to describe models using specific programming languages. Programming-based CAD applications provide multiple benefits to 3D design, but their use remains limited, potentially due to higher entry barriers and extensive programming requirements. Regrettably, a profound lack of understanding of the challenges faced by users of programming-based CAD applications prevents a clear comprehension of the issues of these applications. Furthermore, research addressing CAD challenges has predominantly focused on applications that provide direct manipulation interactions. This doctoral thesis aims to improve the usability of programming-based CAD applications, focusing on their role in Personal Digital Fabrication with 3D printers. Our research seeks to understand and address programming-based CAD users' challenges during the design process. In our first study, we interviewed twenty OpenSCAD users, a leading programming-based CAD application in the 3D printing community. Data analysis via a Reflexive Thematic Analysis (RTA) led to the development of a comprehensive codebook categorizing three main themes: user profiles, 3D design challenges, and 3D printing challenges. Our second study addressed the identified design challenges in linking 3D views with code and difficulties in performing spatial transformations on the model. We proposed to address these difficulties by introducing the concept of bidirectional programming in programming-based CAD, allowing users to interact with both the code and the view. We modified the source code of OpenSCAD to implement this approach, developing bidirectional navigation features and allowing users to edit the model by interacting with the view while the application updates the code coherently. The third study addressed the keystone challenge of defining geometric properties in parametric designs. After analyzing 30 OpenSCAD models, we developed bidirectional programming features in OpenSCAD to facilitate the definition of parametric properties, directly extracting information from the view to use in the code. Experimentation with OpenSCAD users showed our solution may streamline design, reduce errors, and lower entry barriers for newcomers
The Experimental Calibration of a Two-Storey Guarded Hot Box Test Apparatus
No full textA guarded hot box (GHB) test apparatus uses controlled environmental chambers to simulate interior and exterior climate conditions at steady state to assess the thermal performance of building envelope assemblies. Canada’s first two-story GHB was constructed at Carleton University’s Centre for Advanced Building Envelope Research. Prior to using this apparatus for research, a commissioning process was conducted following ASTM C1363. Through experimental procedures, uniform air velocity profiles were established within the environmental chambers with less than 10% variation. The effective thermal resistance of the environmental chamber simulating interior conditions was experimentally determined as 3.24 m2°C/W. The sources of heat loss within the apparatus were quantified by conducting multiple tests over a range of operating conditions using a calibration test sample with known thermal properties. As a result, correction factors were determined to increase the accuracy in assessing the thermal characteristics of building envelope assemblies in standard research operations
A review of motion retargeting techniques for 3D character facial animation
No full textPublished by Elsevier Ltd
Investigating Developer Emotions in Technical Chat Conversations: The Study of Discord and Slack
No full textIn this thesis we aim to address this gap by undertaking a comprehensive exploration of chat conversations within programming communities surrounding Clojure, Python, and Racket, on Discord and Slack. We centre our investigation on chat conversation emotion detection. We have curated and published an annotated dataset of chat conversations, encompassing primary and secondary emotions. Our dataset, CloPyRack, consolidates disentangled chat conversations from the DISCO and SLACK datasets for Clojure, Python, and Racket. We evaluate the performance of existing emotion detection models to assess their efficacy in chat conversation emotion detection. Our evaluation reveals that the pre-trained large language model ALBERT outperforms EMTK models, exhibiting approximately a 68.57% higher F1-score. Moreover, our analysis of the classified datasets for both Discord and Slack unveils a pattern of higher negative emotions, prevalent in Discord
Pattern Engineered Bi-Directional Leaky-Wave Antennas with Independent Beam-Scanning Laws
No full textThree different designs for multi-directional Leaky-Wave Antennas (LWAs) with suppressed stop-bands and frequency scanning capabilities are designed and experimentally verified. The possibility for multi-directional LWAs with independent scanning laws is investigated using bi-directional double slot-pair and edge-fire Vivaldi antennas with different broadside frequencies on each side. Subsequently, two bi-directional edge-fire Vivaldi antenna designs with single- and double-feed lines are proposed. All antenna designs are verified using full-wave simulations before being fabricated and tested. Experimental results for all fabricated antennas are shown, demonstrating good agreement with the simulations. A tri-directional antenna design was also verified using full-wave simulations in Ansys FEM-HFSS. All designed antennas can be engineered to point high gain beams along specified angles, and can be adapted to provide precision control of the independent output beams in various frequency bands
Performance Measurement in the Public Universities in the Province of Ontario, Canada
No full textThe present thesis aimed to explore the use of performance measurement by the public universities in the Province of Ontario, Canada, primarily from institutional theory and, to a lesser degree, contingency theory perspectives. This examination was based on the most salient relationships identified in the conceptual framework developed by the researcher, which shows how organizational internal and external factors contribute to incentivizing the use of performance measurement by Ontario universities and what the entailed potential consequences are for those organizations. An exploratory case study was conducted using an inductive interpretative approach and mainly qualitative research methods. The researcher utilized reflexive thematic analysis, developed by Braun and Clarke (2006), and he adopted a single case study method with a strategically selected group of 11 Ontario universities being treated as embedded units of analysis. The 43 respondents were targeted mainly at the senior managerial levels of universities, which have the potential to impact the utilization of performance information. The study determined that the use of performance measurement in Ontario universities is significantly encouraged by political and regulatory factors, such as strategic mandate agreements, performance-based funding, and academic accreditation bodies. In addition, larger universities have more resources than the smaller ones to implement sophisticated performance measurement systems, while more complex organizations impose the use of performance measurement to a greater extent than the less complex ones. Furthermore, the use of performance measurement has an important contribution in the process of organizational learning and development by using data to identify areas of strength and weakness in organizations. By communicating them to the public, performance data can reveal the organizational contribution to the community and improve organizational accountability, transparency, and legitimacy. In addition, performance information is a main instrument used in comparisons and rankings of universities, which, in turn, impacts institutional public image. However, participants also unveiled some unintended consequences of using performance measurement. For instance, when organizational focus on performance measurement is only on some targeted domains or when performance indicators are poorly selected by universities or other interested organizations, the global improvement of organizational performance can be adversely affected