Theses

Permanent URI for this collectionhttps://uwspace.uwaterloo.ca/handle/10012/6

The theses in UWSpace are publicly accessible unless restricted due to publication or patent pending.

This collection includes a subset of theses submitted by graduates of the University of Waterloo as a partial requirement of a degree program at the Master's or PhD level. It includes all electronically submitted theses. (Electronic submission was optional from 1996 through 2006. Electronic submission became the default submission format in October 2006.)

This collection also includes a subset of UW theses that were scanned through the Theses Canada program. (The subset includes UW PhD theses from 1998 - 2002.)

Browse

Now showing 1 - 20 of 17188

Manila's Other City: Toward a Counter-Relocation Approach
(University of Waterloo, 2025-12-01) Reyes, Mary Angeline
This thesis explores the spatial and economic dynamics of informality in Barangay 105, Tondo, Manila, one of the most densely populated informal settlements in Metro Manila. Located along the industrial edge of the city, the barangay is shaped by rural displacement, state-led resettlement efforts, and infrastructural neglect. At the heart of the site are 25 warehouse structures, originally constructed as temporary relocation housing. Over time, these buildings have been incrementally transformed by residents into permanent live-work spaces, generating a distinct informal morphology that mirrors broader patterns of adaptation across Manila’s socioeconomic landscape. Informality, in this context, is not peripheral but central to the city’s functioning. Approximately 20–35% of Metro Manila’s population resides in informal settlements, many of which operate as self-sufficient ecosystems in the absence of state support. In Barangay 105, waste picking and small-scale recycling form the core of the local economy. Each day, informal workers collect, sort, and resell large volumes of waste, integrating themselves into larger material flows that connect domestic labor to regional and global waste economies. Despite their critical contributions, these workers remain structurally excluded from planning, labor protections, and service provision. To analyze these dynamics, the research draws on large-scale cartography, detailed studies of urban vernaculars, comparative case studies, and the documentation of daily routines. Government housing responses have historically relied on mass relocation, often displacing communities to distant peripheries. Programs such as those led by the National Housing Authority (NHA), the Community Mortgage Program (CMP), and the Zonal Improvement Program (ZIP) have repeatedly failed to address the needs of informal residents, instead severing their access to livelihoods and social networks. This thesis critiques these relocation paradigms and proposes a counter-relocation approach: one that strengthens communities in place rather than removing them. The design focuses on reimagining 15 of the existing warehouse structures as a distributed network of community depots: multi-use infrastructures that embed housing and economic production into the urban fabric. The project centers incremental, solidaristic, and community-led spatial strategies that reflect and strengthen the informal socioeconomic landscape of Tondo.
An Investigation of Overt Visual Attention and Gaze Behaviour in Social Human-Robot Interaction and Human-Computer Interaction Contexts
(University of Waterloo, 2025-12-01) Shaghaghi, Sahand
In human-human and human-robot interaction gaze has a consequential role as a type of non-verbal communication behaviour, affecting the social interaction depending on gaze behaviour's characteristics. As such, gaze behaviour has been a topic of major research throughout the past number of years since a better understanding of gaze behaviour could lead to design of robot behaviour for social interactions. In the context of the human-human interaction (HHI) and human-robot interaction (HRI) studies, gaze behaviour has been seldom investigated while taking into consideration all social interaction elements including interaction partners' personalities and social roles in addition to the social context. There are a number of studies which investigate conversational roles and personality matching in relation to gaze behaviour in the context of HRI in separate studies. However, works which investigate gaze behaviour in tandem with these social interaction elements are needed since such a study will contextualize gaze behaviour in relation to variations in these social elements (e.g. gaze behaviour characteristics based on introverted and extroverted personalities) while taking into consideration the compounded effects of these social elements in combination. What this thesis accomplishes is incorporation of all these social elements in tandem with gaze, all under the umbrella of one body of research. Utilization of this integrative approach was inspired by recent HRI literature, encouraging the investigation of verbal and non-verbal social interaction elements together with social interaction elements. This thesis investigates gaze behaviour in the context of HRI while taking into account social role and designed personalities in robotic platforms. As the social context, this thesis explores dyadic human-robot interactions involving objects of discourse from a gaze-centric point of view while considering the robot's gaze-centric perspective and the participant's gaze-centric perspective. Four major studies are conducted in the context of this thesis to fulfill this exploration. Tools for recording overt visual behaviour are vital in conducting human-computer interaction (HCI) research. However, specific tools enabling the recording of these metrics in online settings, facilitating video viewing were not available, therefore Study 1 created the FocalVid platform. This platform collects cursor location attentional data for the participants in online settings such as Amazon Mechanical Turk. The cursor metrics gathered through this platform were then compared to eye tracking data and our rendition of another relevant platform (BubbleView). It was determined that human gaze and cursor movements are distinct but have similarities in relation to velocities and dwell timing. This platform allowed for large-scale data collection for HCI and HRI studies, which is not possible in the context of in-person studies. Personality and social role are major elements of social interactions; however, perception of designed introverted/extroverted personalities for the humanoid iCub robot were not previously examined and additionally these two elements have not been explored simultaneously in the previous literature involving the iCub robot. In the second study, I explore the participants' perception of a robot in interactions between a robot and a human actor utilizing recorded online scenarios. In this study, the robot takes on different social roles while embodying different personalities. The robot is either a teacher, a student or a collaborator while either introverted or extroverted. To conduct this study, the Amazon Mechanical Turk platform and HRI video recordings were used. I discovered the presence of perceiver effects in participants’ assessment of the robot’s Ten-Item Personality Inventory (TIPI) dimensions perception vs. self TIPI dimensions, where participants' self-assessment of their personality correlated to their assessment of robot’s personality. TIPI questionnaire is a measure used to assess personality dimensions. It was also determined that the designed robot personality was perceived accordingly by the participants. These findings indicated that even though participants’ self-assessment of their personality dimensions affects their perception of the robot, they could still perceive the robot’s designed personality as intended. Observation and analysis of people’s overt visual attention dynamics in HRI could allow for better understanding of these interactions however, such overt attention while considering social interaction elements have not been previously explored in detail. The third study investigated participants' overt visual attention in the context of dyadic social settings using the FocalVid platform. In this study, I was also interested in the efficacy of the use of the FocalVid platform to collect attention metrics relating to such social settings. This study, taking advantage of the HRI scenarios designed in Study 2 and using the FocalVid platform, recorded the cursor attentional data for participants while the robot was enabled with different social roles and personalities. It was determined that the robot’s social role and personality significantly affected the participants’ overt visual attention. It was also determined that the presence of the FocalVid platform did not adversely affect the perception of the robot. Gaze studies in Human Robot Interaction should investigate both the human partner’s gaze behaviour’s effect on the social interaction, in addition to the robot’s gaze behaviour’s effect on the social interaction. A limited number of studies have explored the effects of gaze-architecture-enabled robots' behaviour on social interaction. In the fourth study, after the design of gaze-based interaction architectures based on Social Gaze Space taxonomy in dyadic interactions involving objects of discourse, the effects of using these gaze interaction architectures for robot gaze control were evaluated utilizing eye tracking data and Human Robot Interaction questionnaires. Through this study, it was determined that the SGS-IA architecture led to higher visual engagement by the participants towards the robot’s face and eye region compared to the TutorSpotter architecture, which was used for comparison purposes. One of the main contributions of this thesis is the design and evaluation of these gaze-based interaction architectures for anthropomorphic humanoid robots involved in human-robot interactions. All four of these studies were geared towards gathering a better understanding of gaze behaviour in HRI and HCI. Studies 1 and 2 had a preparatory role to this end. Study 1 allowed us to design the FocalVid platform and to investigate the attention metrics gathered through this platform against gaze metrics in this Human Computer Interaction platform. Study 2 allowed us to design the Human Robot Interaction scenarios needed for Studies 3 and 4. Study 3 investigated gaze behaviour of the human interaction partner involved in Human Robot Interaction using the FocalVid platform, and in Study 4 we designed and evaluated a gaze interaction architecture for the iCub robot through an in-person Human Robot Interaction study. These studies allowed for better understanding of the role of gaze behaviour in social HRI settings. These studies also enabled us to design gaze-specific interaction architectures for the iCub robot.
Rooted Elsewhere: Understanding the impact of immigration on health and wellbeing from the perspective of Black immigrant women in Ontario
(University of Waterloo, 2025-11-28) Ike, Nnenna Arianzu Uma
Introduction Canada has long relied on immigration to shape its demographic landscape, social fabric, and economic development. Historically, immigration policies have favoured white European settlers, but changes to the Canadian Immigration Act opened pathways for more diverse populations, including Black immigrants from the Caribbeans and various African countries. However, Black immigrants, particularly women, continue to face systemic inequities in healthcare, employment, housing, and other areas of life. Black immigrant women often carry the compounded weight of racialized and gendered expectations as they navigate caregiving responsibilities, financial pressures, and resettlement barriers in their new socio-cultural context in Canada. When combined, these social determinant of health can negatively impact their health and resilience as they resettle in Canada. Canadian immigrants are rigorously vetted for the positive educational, health and economic impact they can make to the economy however, their health start to decline after their arrival to the country. Though studies abound establishing patterns linking unmet needs, structural and institutional inequity to immigrant health, for Black immigrant women, health is further shaped by intersecting forces such as race, gender, and migration. While quantitative studies have captured broad patterns in immigrant health, they often fail to capture the lived experiences behind those numbers. This study offers critical insight into the social and emotional complexities of resettlement, health-seeking behaviours, and identity negotiation of immigrants and their impact on physical and mental health. It contributes to a growing body of qualitative scholarship that centres the voices of Black immigrant women in Canada, by offering a deeper understanding of how immigration and resettlement experiences shape their health and well-being. Research Aim This thesis explores the multifaceted immigration and resettlement experiences of Black immigrant women in Ontario, Canada, with a focus on the implications for their physical and mental health. To capture the depth and complexity of the participants’ experiences, this thesis is separated into three papers, each drawing from the same research question but having their own distinct aim. Paper one examines the resettlement challenges specifically encountered by Black immigrant women as they settle and integrate into the Canadian society. It highlights how meeting Canadian immigration eligibility criteria does not ensure effective integration into the Canadian society. It showcases specific resettlement challenges and the resultant impact on their health and wellbeing. Paper two captures the employment experiences of Black immigrant women in the Ontario labour market and produces a three-stage employment narrative common to all participants. Also, paper two highlights the physical and mental health impact associated with the three different stages, along with the different coping strategies deployed by the women. Paper three explores how Black immigrant women make sense of their immigration and resettlement experiences and how the meanings they ascribe to their experiences impact on their health and wellbeing. These meanings significantly influence their choices and behaviors and profoundly impacts their self-identity and engagement with the Canadian society. Methods This thesis employs a qualitative approach to immigrant health research. Purposive sampling recruited a total of twenty-two Black immigrant women living in Ontario, aged 18-54 years. The women participated in virtual and in-person semi-structured interviews that lasted between 45 minutes to two hours. Three analytical approaches were used. First, thematic analysis was utilized in paper one to systematically code and interpret the broader resettlement challenges encountered by participants, such as the persistent devaluation of foreign credentials, experiences of racialized housing discrimination and negative dietary acculturation. Paper two drew on Reissman and Polkinghorne’s narrative analysis approaches to construct a framework that mapped participants’ employment stories across three distinct phases: the initial entry to the labour market, the early employment period, and the longer-term navigation of workplace environment. Paper three used Interpretative Phenomenological Analysis (IPA) to delve into how participants made meaning of their immigration and resettlement experiences within their cultural and socio-political contexts. Findings Findings from the thematic analysis of paper one illuminate how structural and systemic barriers were internalized by the participants, and contribute to their heightened stress, feelings of marginalization, and prolonged sense of frustration which are closely tied to both physical and mental health outcomes. The themes in paper one are: 1.) Hopes and aspiration for a better life, 2.) Facing reality, and 3.) Intentionality. Also, paper one shows that the combination of resilience level and the ability to leverage social capital determine participants’ ability to effectively integrate into the Canadian society. Participants’ employment narratives in paper two expose the emotional strain and disrupted career trajectory experienced by the participants as they tried to continue their professional path after arriving in Canada. Paper two highlights how participants experience structural and social exclusion at the workplace and the resulting negative impact on their physical and mental health. Most importantly, paper two develops a three-stage narrative model of early employment of newcomer Black immigrant women in Ontario. This model links the devaluation of credential and professional experience and racialized gatekeeping of employment to health outcomes in Black immigrant women. Further, this narrative model highlights not only the challenges encountered during the different stages of the employment journey but also how these experiences can contribute to the onset and progression of imposter syndrome (IS). The Interpretive phenomenological Analysis (IPA) method used in paper three foregrounds the embodied and affective dimensions of resettlement by showing how the participants’ identities, aspirations, and understandings of their health evolved in response to both visible and invisible pressures from their immigration. Paper three shows how participants sense of immigration was different before and after their arrival in Canada. Further, it shows that how they internalised and made sense of their experiences determined the decisions and actions they take which in turn impacted on their health and wellbeing. By privileging participants sense-making, paper three reveals the Black immigrant woman’s nuanced portrait of resilience and negotiation of self identity in their destination country. Conclusion Paper one highlights the feelings of disadvantage and stress expressed by all the participants as a result of their resettlement experiences of dietary acculturation, housing and healthcare discrimination. It showcases the results of the intersectionality of the participants identities and how social capital influences the extent to which participants are were able to navigate their resettlement challenges. Paper two portrays how people internalise the devaluation of their educational and professional qualifications because of systemic barriers and sexism and racism in Canadian workplaces. Further, Paper two identifies how participants experience of the imposter syndrome has far-reaching effect on their physical and mental health. Also, it shows that somatic symptoms can be easily developed as a result of workplace stress, and if they are misdiagnosed or overlooked, can lead to more serious and chronic health conditions. Paper three shows that beyond physical relocation, immigration is a deeply emotional and psychological journey shaped by the intersecting forces of gender, race and identity. Participants’ experience of racial trauma and systemic exclusion resulted in a sense of resignation to conserve their emotional energy. However, they demonstrate resilience through positive reframing and biographical reinvention to reconstruct their self identity in order to live and thrive in Canada as an immigrant. Overall, the contributions of this thesis include: 1. A methodological contribution showing productive complementarity between Interpretative Phenomenological Analysis Narrative Inquiry, and Thematic Analysis for capturing both idiographic depth and substantive patterns across the dataset. 2. A theoretical refinement by demonstrating how the Intersectionality lens, backed by the Critical Race Theory, and Migration and Integration theories can be used to capture and categorize data for analytic navigation and story telling. 3. An empirical contribution by way of a three-stage narrative employment model for newcomer Black immigrant women in Ontario that links credential devaluation and racialized gatekeeping to health impacts. Ultimately, this thesis affirms that immigration and resettlement, for Black immigrant women, is not just about relocating and fitting in a new country but is about renegotiating their entire self-identity in environments with limited socio-cultural and professional support. When these are not appropriately addressed, will impact negatively on their health and wellbeing in their destination country.
Safety and Security of Reinforcement Learning for Autonomous Driving
(University of Waterloo, 2025-11-27) Lohrasbi, Saeedeh
In the context of autonomous driving, reinforcement learning (RL) presents a powerful paradigm: agents capable of learning to drive efficiently in unseen situations through experience. However, this promise is shadowed by a fundamental concern—how can we entrust decision-making to agents that rely on trial-and-error learning in safety-critical environments where errors may carry severe consequences? This thesis advances a step toward resolving this dilemma by integrating three foundational pillars: adversarial robustness, simulation realism, and model-based safety. We begin with a comprehensive survey of adversarial attacks and corresponding defences within the domains of deep learning (DL) and deep reinforcement learning (DRL) for autonomous vehicles. This survey reveals the porous boundary between safety and security—both natural disturbances and adversarial perturbations can destabilize learned policies. Motivated by this insight, we introduce the Optimism Induction Attack (OIA), a novel adversarial technique that manipulates an RL agent’s perception of safety, causing it to act with unwarranted confidence in hazardous situations. Evaluated in the context of an Adaptive Cruise Control (ACC) task, the OIA significantly impairs policy performance, exposing critical vulnerabilities in state-of-the-art RL algorithms. To counter the demonstrated threats, we present a systematic defence architecture. We develop REVEAL, a high-fidelity simulation framework designed to support the training and evaluation of safe RL agents under realistic vehicle dynamics, traffic scenarios, and adversarial conditions. By narrowing the gap between abstract simulation and real-world complexity, REVEAL facilitates rigorous and nuanced testing, which is essential for safety-critical applications. To enhance learning efficiency within this environment, we employ a transfer learning (TL) strategy: policies initially trained in simplified simulators (e.g., SUMO) are adapted and fine-tuned in REVEAL, leading to faster convergence and improved safety performance during both training and deployment. Central to our approach is the development of a Multi-Output Control Barrier Function (MO-CBF), which simultaneously supervises throttle and brake commands to enforce safety constraints in real time. Rather than relying on hard overrides, the MO-CBF operates cooperatively with the learning agent—gently adjusting unsafe actions and introducing corresponding penalties during training. This enables the agent not only to learn safe behaviour but also to internalize safety principles and anticipate potentially unsafe scenarios. Our empirical evaluation demonstrates the effectiveness of the proposed framework across a spectrum of disturbances, adversarial inputs, and realistic high-risk maneuvers. The results consistently show improved safety and robustness, highlighting the framework’s capacity to transform RL agents from vulnerable learners into trustworthy autonomous systems. In summary, this thesis presents a comprehensive methodology for safe and secure RL in autonomous driving. By grounding agent training in high-fidelity simulation, leveraging adversarial awareness, and embedding real-time model-based safety mechanisms, we provide a cohesive and scalable pathway toward deploying RL in the real world with confidence.
How Architectural Style, Height, and Complexity Influence Perceived Oppressiveness in Urban Spaces
(University of Waterloo, 2025-11-27) Lapietra Garcia, Thomas
The design of urban environments strongly influences psychological experience, yet research on how building form influences affective responses remains limited. This study used immersive virtual reality to examine the combined effects of architectural style (modern vs. contemporary), building height (low-, mid-, and high-rise), and façade complexity (low, medium, high) on affective perceptions of urban streetscapes. Forty-nine participants explored 18 virtual environments and rated each on oppressiveness, openness, restoration, arousal, and environmental liking. Results showed that greater building height consistently increased perceived oppressiveness and arousal while reducing openness, stress restoration, and liking. Greater façade complexity increased preference, openness, and restoration, and buffered the oppressive effects of high-rises, particularly in modern-style settings. Participants also expressed a clear preference for low- and mid-rise settings over high-rises. These findings reiterate and expand on the restorative and aesthetic benefits of architectural complexity and the value of human-scale design in supporting psychological well-being in urban dwellings.
Quantifying Endplate Deflection in Response to Cyclic Load Exposures Using a Porcine Cervical Spine Model
(University of Waterloo, 2025-11-27) Watson, Michael
The vertebral endplate is a thin layer of cartilage and bone that separates the intervertebral disc from adjacent vertebral bodies and facilitates the transmission of compressive force through the spine. Despite this essential function, it remains the weakest component of the vertebra-disc unit and is highly susceptible to mechanical failure. Endplate failure typically arises from localized tensile strains that manifest as deflection, defined as the out-of-plane displacement of the surface under load. While prior work has demonstrated inferior endplates of intervertebral joints exhibiting greater deflection and higher incidence of failure than their superior counterpart, current techniques for quantifying endplate deflection face notable limitations. Early studies using metallic markers or displacement transducers required drilling channels into the vertebral body, potentially exaggerating deformation by weakening subchondral bone support. Imaging-based approaches, particularly micro-CT, offer high spatial resolution but are limited to static or stepwise loading due to temporal constraints. These static conditions do not capture the cyclic loading patterns experienced by the spine during daily activity, where repeated deformation can cause fatigue-induced microdamage and eventual failure. Additionally, static loading promotes excess fluid loss from the nucleus pulposus, altering endplate deflections in ways that do not reflect physiological motion. Consequently, existing measurement techniques may misrepresent true endplate behavior and are unable to evaluate changes in deflection as a function of cyclic load exposure. This study addresses these limitations by developing a unique method to assess endplate deflection during cyclic loading without requiring prolonged stepwise protocols or causing damage to the vertebral bone. By comparing superior and inferior endplates across different load magnitudes and cyclic durations, this work aims to clarify the mechanisms underlying endplate vulnerability and further validate the porcine cervical spine as an experimental model for human lumbar spine deflection. Eighteen porcine cervical spine functional units (C3C4, C4C5, and C5C6; n = 6 per level) were dissected to yield 36 individual vertebrae. High-resolution laser profilometry was then used to capture the topography of the caudal endplates of C3, C4, and C5 and the cranial endplates of C4, C5, and C6. Custom indenters, designed as negative molds of the nucleus-occupying endplate region, were created from the resulting surface scans and fabricated via 3D printing. Specimens were then oriented such that the tested endplate was in a neutral position and subjected to a normalized haversine waveform, ranging from 0.3 kN to 30% of the predicted ultimate compressive strength using a servohydraulic materials testing system. The cycle-dependent changes in endplate deflection were measured at 0, 1000, 3000, and 5000 total cycles. At each time point, endplate deflection measurements were captured via the indenter’s displacement while specimens were exposed to a brief static force of 0.3 kN, 1 kN, and 3 kN, totaling 12 measurements per vertebra. Three separate linear mixed effects models were used to evaluate the impact of loading magnitude, loading cycles, endplate level and the proportion of the nucleus occupying endplate area on superior and inferior endplate deflection within each joint. A fourth linear mixed effects model was used to evaluate the impact of loading magnitude, loading cycles, and joint level on the magnitude of the differences between superior and inferior endplate deflection. Utilizing this novel methodology, this study was the first to quantify endplate deflection under cyclic loading conditions, observing greater deflection of the inferior endplate across all spinal levels, except at baseline (0.3 kN, 0 cycles). This method also enabled comparison of deflection rates between endplates, with the C4C5 and C5C6 inferior endplates showing a significantly greater rate of deflection during the first 1000 cycles. Among joints, C4C5 exhibited the largest difference in superior and inferior endplate deflection compared to C3C4 and C5C6. Endplate deflection was not influenced by the proportion of the nucleus occupying endplate area at any spinal level. Lastly, as the first study to examine endplate deflection in porcine cervical vertebrae, the observation of greater inferior endplate deflection being consistent with human cadaveric studies further supports the validity of this model. Overall, this study demonstrates the utility of a novel methodology for measuring and comparing superior and inferior endplate deflection under cyclic loading.
Zero-Knowledge Proof-Enabled SAT Co-processor for Blockchain Systems
(University of Waterloo, 2025-11-26) Yusiuk, Vladyslav
This thesis explores the possibility of building classical SAT solvers in Circom Domain Specific Language to create zero-knowledge proofs (ZKPs) usable in blockchain contexts. I implemented DPLL and Chaff as arithmetic circuits within Circom and analyze them based on constraint count, proving delay, and zk-SNARK verification layers. With this evaluation, the aim is to determine the feasibility of solvers integration into off-chain computation systems and rollup-centric architectures on Ethereum. The findings indicate that incorporating SAT solvers within zero-knowledge circuits is achievable though some degradation in efficiency occurs based on algorithm used and input representation. This research provides a thorough assessment of known SAT methods across an unconventional boundary, linking symbolic logic with blockchain technologies reliant on zk-SNARKs.
Development of High Strength Aluminum Alloys for Directed Energy Deposition Additive Manufacturing
(University of Waterloo, 2025-11-26) Waqar, Taha
Among additive manufacturing (AM) techniques, directed energy deposition (DED) is of particular interest for structural Al alloys, as it combines the faster cooling rates with the flexibility to repair or build large-scale geometries. The localized thermal cycling inherent to the DED process influences solidification behavior, grain refinement and precipitate evolution for high strength age-hardenable Al alloys such as Al 7075, which in turn governs the mechanical performance. These capabilities position DED as a promising pathway for expanding use of high strength heat-treatable aluminum alloys in aerospace and automotive applications where a good strength to weight ratio is crucial. However, Al 7075 tends to crack during solidification and possesses a limited service temperature range. The research conducted explores the tailoring of an existing Al 7075 composition and delves into the development of novel Al alloy compositions for DED-AM processes. In the initial phase of the research, laser-directed energy deposition of Al 7075 wire feedstock enhanced with TiC nanoparticles to promote grain refinement was investigated. It was found that the combination of high laser power (3400 W) along with low travel speed (400 mm/min) and low wire feed speed (400 mm/min) resulted in the reduction of lack of fusion defects and reducing cracks within the multilayer prints. However, substantial evaporation during printing led to a reduced amount of Mg and Zn bearing phases in the as-printed samples. It was shown that the direct aged sample heated for 5 hours was of comparable hardness to the T6 (solution heat treated and then artificially aged) sample (115 HV0.5), which highlights the presence of solute trapping in the as-printed material. To compare the behavior of the same Al 7075 + TiC wire feedstock under arc-based solidification conditions, the research continued to investigate the microstructural evolution and mechanical response of Al 7075 reinforced with TiC nanoparticles processed via arc-based DED, with a particular focus on aging behavior. Grain refinement was primarily attributed to heterogeneous nucleation and grain boundary pinning by TiC clusters. Moreover, TiC inoculants influenced solute redistribution, driving segregation of Mg and Cr, which in turn altered the precipitation behavior during aging. Heat-treated samples revealed the co-formation of MgZn₂ strengthening precipitates and the E-phase (Al18Mg3Cr2), with the latter contributing to the heterogeneous distribution of precipitates. These findings highlight both the benefits and challenges of TiC inoculation in tailoring microstructure and age-hardening response in arc-DED processed Al 7075 alloys. The second phase of the research presents the design and evaluation of a novel Al-Ce-Mg alloy tailored for wire arc-DED. The objective was to overcome the limitations of conventional high-strength aluminum alloys, which suffer from solidification cracking, volatile element loss, and poor thermal stability at elevated temperatures. Alloy selection was guided by CALPHAD simulations, leading to the identification of a near-eutectic Al-10Ce-9Mg composition. Thin-wall structures were fabricated, and porosity was quantified using micro-computed tomography, supported by high-speed imaging that revealed oxide-film entrapment as the dominant cause of porosity. The solidified microstructure consisted of α-Al, eutectic, and primary Al₁₁Ce₃ phases, as well as β-AlMg phase, which contributed to both strength and thermal stability. Compression testing demonstrated high room-temperature strength but brittle failure. At elevated temperatures, however, the alloy retained superior strength compared to conventional precipitation-strengthened Al 7075 alloy, even after extended thermal exposure. This observation was attributed to the stability of Al-Ce intermetallics. Incorporation of Sc into Al-Ce-Mg alloys can provide a dual strengthening and thermal stabilizing effect. Therefore, in the final phase of the conducted research, an Al-8Ce-8Mg-0.2Sc alloy was developed. Laser surface remelting was employed to replicate AM-like conditions, producing a refined bimodal grain structure and fragmenting coarse Al₁₁Ce₃ networks into discontinuous, blocky morphologies. Compared to the as-cast state, the remelted alloy exhibited increased hardness (114.5 HV1 vs 133 HV1), aided by refined grains and secondary phases such as Al11Ce3 and Mg2Si. Direct aging produced an irregular hardening response, with peak hardness achieved at 375 °C for 1 h due to the precipitation of coherent Al₃Sc nanoprecipitates. Long-term thermal exposure at 200 °C for up to 1000 hours showed negligible hardness loss and minimal coarsening of Ce-bearing intermetallics. Strengthening contributions were dominated by Al₃Sc precipitation, supported by solid-solution, grain refinement, dislocation hardening, and stable Al₁₁Ce₃ dispersoids.
Heterogeneous Decomposition of Convolutional Neural Networks Using Tucker Decomposition
(University of Waterloo, 2025-11-26) Mokadem, Frank
Convolutional Neural Network (CNN) remain the architecture of choice for computer vision tasks on compute-constrained platforms such as edge and personal devices, delivering both close to state-of-the-art performance metrics and linear inference complexity with respect to input resolution and number of channels. However, the deployment of larger and more complex CNN architectures is limited by the restrained memory offered by such platforms. This brings about a need to compress pretrained CNN into smaller models in number of parameters while controlling for degradation in performance. This thesis tackles CNN compression using low rank approximation of convolution layers using Tucker Decomposition (TD). We introduce a new heuristics-based Neural Architectural Search procedure to select low rank configurations for the convolution tensors, which we call Heterogeneous Tucker Decomposition (HTD). Standard low rank approximation using TD factorizes and approximates convolution layers using uniform ranks for all convolution tensors, then applies a few fine–tuning epochs to recover degradation in performance. An approach we show to be suboptimal against a heterogeneous selection of ranks for each convolution layer, followed by same number of fine-tuning epochs. Our primary contribution is the development and evaluation of TD, which applies layers-pecific compression rate (low rank divided by full rank) inferred from a Neural Architectural Search (NAS) process. Furthermore, we introduce a sampling heuristic to efficiently explore the search space of layer-specific compression rates, thus preserving performance while significantly reducing search time. We present a mathematical formulation for the HTD optimization problem and an NAS algorithm to find admissible solutions. We test our approach on multiple varieties of CNN architectures: AlexNet, VGG16, and ResNet18, adapted for the MNIST classification task. Our findings confirm that HTD performs better than TD on all models tested. For the same compression rate, HTD enables to recover a higher precision after fine-tuning, with gains ranging from 1.2% to 5.8%. For equivalent accuracy targets, HTD delivers 15-30% higher compression rates than TD. This thesis advances Neural Architectural Search by highlighting the efficacy of heterogeneous tensor decomposition approaches. It provides a robust framework for their implementation and evaluation, with significant implications for deploying convolutional deep learning models in resource-limited settings. Future work will explore incorporating low-rank constraints as a regularization objective during training, potentially enabling end-to-end compression-aware optimization.
Proactive Characterization of Wildfire Impacts on Drinking Water Treatability
(University of Waterloo, 2025-11-26) Bahramian, Soosan
Forested catchments are important sources of drinking water globally. They are increasingly threatened by disturbances, prominently climate shocks, including large wildfires. Wildfires alter watershed hydrology and biogeochemistry, leading to reduced infiltration, increased overland flow, and enhanced delivery of sediments, burned vegetation, and pyrogenic material into aquatic systems. Such inputs can alter drinking water source quality and challenge treatability. Ash is the residual material from wildland fuel combustion, composed of mineral particles and organic matter that can leach into water. While inorganic dissolved compounds from ash can impact water quality by, for example, increasing ionic strength and alkalinity, water-extractable organic matter (WEOM) from wildfire ash contributes to increased post-fire dissolved organic carbon (DOC) concentrations. During drinking water treatment, higher DOC concentrations increase chemical demand (e.g., coagulant, disinfectant), enhance the formation of potentially harmful disinfection by-products (DBPs), cause taste and odor issues, and promote bacterial regrowth in distribution systems. These impacts may also necessitate new infrastructure to manage changes in source water quality, ultimately increasing overall treatment costs. Although they cannot reflect all watershed processes, bench-scale evaluations provide valuable insights into wildfire impacts on drinking water treatability by isolating treatment-relevant mechanisms at controlled laboratory conditions. However, different approaches used to prepare wildfire ash-impacted waters (WAIWs) limit the inferences that can be drawn from them. Here, key factors (e.g., mixing duration and condition, ash-to-water ratio, and source water quality) that can impact organic matter leaching from wildfire ash to water were investigated. WEOM concentration increased within the first 24 hours of mixing before plateauing or declining as mixing progressed, regardless of ash type and background water source. Continuous mixing yielded higher WEOM concentrations than stagnant conditions, indicating that particle-particle interactions and surface exposure enhanced leaching. WEOM yield also decreased as ash-to-water ratios increased. Despite anecdotal suggestions, a relationship between wildfire ash color and WEOM concentration was not observed (Chapter 2). Wildfire ash collection methods may also impact inferences drawn from bench-scale drinking water treatability assessments. Unburned vegetation, rocks, or other debris may have physico-chemical properties different from those of ash deposits; thus, increasing uncertainty in treatability assessments. Dry ash homogenization methods (i.e., manual separation, sieving, and pulverization) were investigated because they may mitigate these impacts. Sieving was shown to be the most practical and reliable method for ensuring ash homogeneity. Pulverization enhanced organic matter release from large particles by increasing surface area, but it also generated aerosolized ash, complicating sample handling. In addition, pulverization altered WEOM character, potentially by increasing the availability of smaller organic matter compounds previously encapsulated within ash particles or by mechanically fragmenting larger organic molecules into smaller compounds (Chapter 3). Subsequent investigations examined the role of settleable ash solids (SAS), a previously overlooked fraction of wildfire ash. SAS substantially increased water alkalinity and make pH control for coagulation extremely difficult. Although pH adjustment enhances DOC removal from WAIW, SAS increased acid demand substantially. The removal of SAS reduced both alkalinity and acid demand; however, as ionic strength was concurrently reduced, floc formation and turbidity reduction for a given coagulant dose decreased somewhat. A limited complementary analysis was conducted to evaluate whether atmospheric ash deposition could also act as a significant driver of source water quality and treatability change. While the impact of atmospheric deposition of ash on water alkalinity depends on the surface area of water body, only exceptionally high atmospheric ash loading could meaningfully alter source water alkalinity in reservoirs that hold large volume of water (Chapter 4). Wildfire ash alters multiple aspects of water quality concurrently, including turbidity, DOC concentration and character, and alkalinity, so its overall implications for water treatment cannot be adequately assessed by examining individual mechanisms in isolation. Coagulation experiments with WAIWs demonstrated these interacting impacts. At low coagulant (i.e., alum) doses, turbidity was effectively reduced, yet DOC removal remained limited, despite pH adjustment to coagulant-specific optima. Enhanced coagulation combined with higher alum doses improved DOC removal but introduced trade-offs, as turbidity reduction declined somewhat because of reduced ionic strength associated with decreased alkalinity. The results indicated, while wildfire ash can severely deteriorate water quality by increasing turbidity, alkalinity, DOC concentration, and aromaticity, which may increase coagulant demand or necessitate more advanced treatment methods, the underlying coagulation mechanisms for WAIW remain consistent with those in natural waters. Thus, wildfire ash does not present fundamentally new challenges to coagulation; rather, the magnitude of water quality changes following wildfire can pose risk to treatment performance and operational resilience (Chapter 5). Collectively, this research demonstrates that while bench-scale studies cannot fully replicate the complexity of post-fire watershed processes and wildfire impacts on water quality, they remain essential for isolating and investigating the specific effects of wildfire ash on drinking water treatment processes. Accordingly, it is practical to adopt methods that maximize the extraction of organic matter from wildfire ash and represent worst-case treatment scenarios. These methodological insights help ensure the comparability of bench-scale investigations. This work also shows that wildfire impacts coagulation primarily by complicating pH control and deteriorating drinking water source quality, increasing the need for more intensive treatment processes. Overall, this research establishes a robust methodological foundation for reliably assessing wildfire ash impacts on water quality and for informing the development of strategies to mitigate wildfire impacts on drinking water treatability.
Design and Implementation of a Robust State of Charge Estimation Approach for a Single Battery Cell, a Hardware-in-the-Loop Test Bench, and a Battery Disconnect Unit for an Electric Vehicle Battery Pack
(University of Waterloo, 2025-11-25) Pham, Nguyen Truong Son
As transportation electrification accelerates, battery-powered vehicles, including cars, airplanes, and boats, are rapidly emerging. This thesis provides solutions and practical insights on two key topics: implementing robust machine learning algorithms on commercial Battery Management System (BMS), and building a high-performance Battery Disconnect Unit (BDU). The experience was gained during participation in the North American Battery Workforce Challenge. First, two machine learning approaches for State of Charge (SoC) estimation are introduced. The first approach is an adaptive algorithm using SoC-OCV-T (State of Charge-Open Circuit Voltage-Temperature) lookup table and Extreme Learning Machine (ELM). The experiment began at 100% SoC, with temperature ranging from -20°C to 60°C. From -20°C to 0°C, the maximum absolute error (MAE) ranged from 0.030 to 0.025. In the mid-range from 5°C to 40°C, the MAE decreased to within 0.015 to 0.020 range. Lastly, at higher temperature range of 45°C to 60°C, the MAE was below 0.013. In the second approach, advanced differential features are added to improve the accuracy of the ELM model, particularly below 0°C. Under noisy condition, both the maximum absolute error (MAE) and the root mean square error (RMSE) were reduced to below 1.5% at -20, 20, and 60°C. Both algorithms were validated on a customized Hardware-in-the-loop (HIL) test bench. The HIL platform was developed to streamline validation of algorithms such as SoC estimation. Finally, the thesis details the design and testing process for the BDU, highlighting key design considerations, test results, and engineering challenges.
Robust Nonparametric Inference on Manifold Spaces
(University of Waterloo, 2025-11-25) Mozaffari, Ahmad
We propose rank-based procedures for robust and nonparametric statistical inference on manifold spaces. Particularly, we focus on the problems of multi-sample hypothesis testing, multiple change point analysis, and statistical process monitoring when data lie on a Riemannian manifold. These methodologies provide a unified framework to deal with various types of data structures such as matrices, curves, surfaces, networks, to name a few. These types of datasets frequently appear in a broad set of applications such as communication networks, manufacturing, computer vision, autonomous systems and robotics. We evaluate the proposed methods considering various types of object data such as matrices, curves, text mining data, networks, shape data and landmarks. In Chapter 2, we develop robust and nonparametric methods for hypothesis testing when data lie on a manifold. We demonstrate that ranks generated from data depth can be used for two-sample and multiple sample hypothesis testing of change in location and scale parameters. Several important properties of these tests such as asymptotic convergence, size and power, robustness with respect to qualitative-robustness and breakdown point are developed under mild nonparametric assumptions. These tests have several advantages, they have a simple distribution under null, they are computationally cheap, and they enjoy invariance properties. We demonstrate the efficacy of these methods with a numerical simulation and a data analysis. We show that these tests are robust when data are heavy tailed or skewed, and have higher power compared to their competitors in some situations, while still maintaining a reasonable size. In Chapter 3, we propose robust and nonparametric single and multiple change point detection methods for stochastic processes defined on manifolds. These methods consider a variant of CUSUM statistic which is based on the rank of data depth. We demonstrate that changes in the rank of depth values can be used to detect change in the distribution of data lie on manifolds. To detect more than one change point, we consider binary segmentation and wild binary segmentation algorithms along with the proposed data depth rank CUSUM statistic. We demonstrate that both of these algorithms are consistent estimators of the number of change point(s) and the location of change point(s). In addition to asymptotic results, we develop nonasymptotic sharp bounds for single and multiple change point estimators. These test statistics can be applied to both intrinsic and extrinsic manifold analysis frameworks. In simulation, we compare our methods against several methods from the literature, and demonstrate that the proposed methods outperform their competitors in some situations where dataset is contaminated with outliers. We also present the application of our methods to vehicle health monitoring, traffic monitoring on highways, and mall pedestrian surveillance. In Chapter 4, we extend these methods to the setting of statistical process monitoring. We investigate statistical process monitoring scheme on general metric spaces, and propose exponentially weighted moving average, CUSUM, and Mann-Whitney moving average Shewhart control charts using rank of data depth. These methods are nonparametric and robust to outliers through the use of data depth ranks. We show that when sample size is large, our methods have simple behaviour under the null hypothesis. Since our methods are based on data depth ranks, we do not need the estimate of covariance operator which makes our method computationally cheap. Such advantages make these methods a favorable choice for online process monitoring. We demonstrate the robustness of these methods theoretically and numerically. We extract several nonparametric control charts from the literature for comparative study. Simulation results indicated that the proposed methods outperform their competitors in many situations in terms of out-of-control average run length, while keeping the in-control average run length at a reasonable level. We present the application of our methods to laser power-bed fusion additive manufacturing process. In Chapter 5, we present some possible directions for future research related to dynamic network and longitudinal data analysis on Riemannian manifolds. It is anticipated that the contributions achieved in this thesis will be applicable to a wide range of interdisciplinary research problems.
Towards Honest, Practicable and Efficient Private Learning
(University of Waterloo, 2025-11-24) Mohapatra, Shubhankar
Protecting our personal information is a major challenge in today's data-driven world. When scientists and companies analyze large datasets, they need a way to ensure our individual privacy isn't compromised. This thesis focuses on Differential Privacy, a powerful, mathematical guarantee that places a strict, verifiable limit on how much personal information can be leaked, even if an attacker has the worst-case advantage. Researchers have developed various sophisticated algorithms to accomplish useful tasks, like building machine learning models or generating realistic synthetic data, while maintaining Differential Privacy. Crucially, these operations must be conducted within a predetermined, strict limit, often referred to as the "privacy budget." This budget mathematically quantifies the total acceptable loss of privacy for the entire process, enforcing a crucial trade-off between data utility and individual protection. All routine procedures of the machine learning pipeline, including data cleaning, hyperparameter tuning, and model training, must be performed within the budget. Several tools can perform these tasks in disjunction when the dataset is non-private. However, these tools do not translate easily to differential privacy and often do not consider the cumulative privacy costs. In this thesis, we explore various pragmatic problems that a data science practitioner may face when deploying a differentially private learning framework from data collection to model training. In particular, we are interested in real-world data quality problems, such as missing data, inconsistent data, and incorrectly labeled data, as well as machine learning pipeline requirements, including hyperparameter tuning. We envision building a general-purpose private learning framework that can handle real data as input and can be used in learning tasks such as generating a highly accurate private machine learning model or creating a synthetic version of the dataset with end-to-end differential privacy guarantees. We envision our work will make differentially private learning more accessible to data science practitioners and easily deployable in day-to-day applications.
Decolonizing Disability: access without erasure
(University of Waterloo, 2025-11-24) Musa, Kenyo
This thesis rethinks disability in the Global South by turning to Nigerian open-air markets, rather than institutional settings as primary sites of inquiry. More than points of exchange, these markets are cultural and civic arenas where economic activity intersects with social connection, mutual care, and collective identity. Marketplaces often function as “third places,” sustaining relationships, preserving communal memory, and hosting the negotiation of public life alongside commerce. Within this context, disability is framed not as a fixed biological deficit but as a condition shaped by environments, social structures, and cultural narratives. Drawing on critical disability studies, African epistemologies, and the concept of relational access, the project positions design as a continual negotiation between bodies, space, and practices of care, challenging functionalist approaches that reduce access to technical compliance. A central critique advanced in this research is the co-option of accessibility language to legitimize exclusionary development. In postcolonial African cities, modernization projects often promise accessible infrastructure while simultaneously displacing those most reliant on markets for survival. Under the banner of “ultra-modern” shopping complexes, elderly traders, people with impairments, and low-income groups are frequently priced out, excluded from decision-making, and stripped of long-standing spatial and economic networks. In such cases, access becomes a rhetorical tool for privatization and displacement rather than a pathway to justice. This thesis argues that genuine access must go beyond token infrastructural features to address the deeper social, economic, and political systems that sustain participation. Methodologically, the study combines critical literature with graphical anthropology, using mapping and diagramming to interpret the spatial conditions of Nigerian markets. This approach, informed by Jos Boys’s “Having a Body” framework, highlights how non-normative bodies engage space, revealing barriers such as uneven ground, sensory overload, or disorientation, alongside supports like shared seating, mutual caregiving, and assistance from load carriers. Through this iterative method, the research develops strategies grounded in lived realities rather than abstract standards, emphasizing collective arrangements that sustain participation. The design proposal focuses on Jos Main Market, a once-celebrated hub now in disrepair after arson and neglect. The intervention introduces a spine that organizes utilities and circulation while embedding care nodes for prayer, rest, sanitation, and basic medical support. A market workshop provides space for repair, fabrication, and low-cost assistive devices, affirming resourcefulness and local skill as vital forms of access. At its center, a market plaza serves as a commons, enhancing visibility and offering social services such as collective childcare, community kitchens, thrift collectives, and meeting areas. Together, these spaces strengthen support networks and ensure vulnerable groups remain active within the civic life of the market. Ultimately, the thesis positions open-air markets as sites that resist the misuse of accessibility rhetoric by grounding access in reciprocity and collective care. Rather than treating informality as disorder to be erased, it demonstrates how markets themselves model alternative approaches to spatial justice. By centering lived experience, this project advances a decolonial vision of disability design, one where access is relational, negotiated, and inseparable from economic survival and community life.
Monitoring risk from contaminant mixtures in stormwater with water quality measurements, bioassays, and bioassessment
(University of Waterloo, 2025-11-20) Izma, Gab
Urban stormwater management ponds (SWPs) are increasingly valued not only for their role in mitigating runoff but also for the biodiversity they support in densely developed environments. However, these systems receive complex contaminant mixtures from urban runoff, including pesticides, pharmaceuticals, industrial chemicals, and metals. These pollutants can accumulate in biologically active compartments like biofilms, posing risks that are not always captured by traditional water-based monitoring. My thesis investigates the nature, accumulation, and ecological effects of contaminants in SWPs using a combination of chemical, biological, and toxicological approaches. The objectives of my research were to: (1) characterize pesticide contamination in SWPs using water, biofilm, and passive samplers; (2) quantify pesticide accumulation in biofilms and identify influencing factors; (3) assess the toxicity of contaminated biofilms through dietary exposure; (4) survey the broader suite of urban contaminants in SWPs to develop a stormwater contaminant signature; and (5) examine relationships between environmental conditions and aquatic community composition. In Chapter 2, I surveyed 21 SWPs in Brampton, Ontario for pesticide contamination. I compared three monitoring approaches across the ponds - time-integrated water sampling, biofilm cultured on artificial substrates, and organic-diffusive gradients in thin films (o-DGT) passive samplers - finding that o-DGTs had the highest pesticide detection rates. However, issues with reproducibility in passive sampler data highlighted the challenges of using them for quantitative risk assessment. Despite generally low concentrations in water and biofilm samples, the widespread detection of diverse pesticide classes across all three matrices emphasized the chronic, mixture-based exposures in these ponds and informed recommendations for future monitoring strategies. In Chapter 3, I further investigated the use of biofilms as a sensitive and ecologically relevant matrix for contaminant monitoring. Examining a wider set of pesticide analytes, I found that over half of the pesticides detected in biofilm samples were not detected in water, suggesting that conventional sampling approaches may overlook important alternative exposure routes. Calculated bioconcentration factors (BCFs) varied widely and were not well explained by pesticide properties or water quality variables, pointing to the complexity of contaminant uptake mechanisms. To test the potential toxicity of these contaminated biofilm samples, in Chapter 4 I conducted a series of dietary exposure assays with two invertebrate grazers. Mayfly nymphs (Neocloeon triangulifer) and juvenile freshwater snails (Planorbella pilsbryi) fed with contaminated biofilms from the SWPs showed reduced survival and growth compared to controls. Although the test results did not always correlate with measured pesticide levels, these results support the ecological relevance of biofilm-mediated exposure and suggest the presence of additional stressors not captured in targeted chemical analyses. I further expanded the chemical scope in Chapter 5 by analyzing over 700 unique urban contaminants across water, biofilm, and o-DGT samples. In total, 200 organic compounds were detected, including personal care products and traffic-related pollutants, as well as persistent elevated levels of fecal indicators and chloride. From these data, I developed the Urban Stormwater Contaminant Signature (USCS): a proposed list of common, environmentally relevant compounds to guide future monitoring and toxicity testing in urban aquatic systems. Finally, in Chapter 6 I examine how environmental variables shape aquatic community composition. Diatom and macroinvertebrate assemblages sampled from the SWPs were dominated by pollution-tolerant taxa, with diatoms responding primarily to water quality (e.g., nutrients, chloride, herbicides) and macroinvertebrates more sensitive to habitat features associated with pond naturalization. Landscape-scale metrics (e.g., impervious cover) calculated from buffer zones had limited predictive power, suggesting that local conditions and upstream drainage characteristics play a stronger role in shaping biological communities. This research highlights the need to expand contaminant monitoring in stormwater systems beyond conventional water sampling, incorporating matrices like biofilm and tools such as passive samplers to better reflect the complexities of exposures in urban environments. The detection of numerous unmonitored or rarely assessed compounds suggests that current regulatory frameworks may underestimate the complexity and risk of urban chemical mixtures. Recognizing stormwater ponds as both infrastructure and ecosystems calls for more ecologically grounded approaches to design, management, and risk assessment; ones that support biodiversity alongside water quality improvement and flood protection.
Development of Functional Binders and Li2S@Carbon Nanocomposites for High-Performance Lithium Sulfide Batteries
(University of Waterloo, 2025-11-20) Huang, Zhe
Lithium sulfide (Li2S) is a promising cathode material for lithium-sulfur batteries (LSBs) owing to its high theoretical capacity (1166 mA h g-1) and potential for safer, scalable battery architectures. In contrast to sulfur cathode, Li2S enables direct pairing with commercial anode materials, avoiding the safety risks of lithium metal. Despite these merits, practical application of Li2S is challenged by its hygroscopic nature, which forms insulating LiOH/Li2O surface layers that cause a large first-charge overpotential; its high melting point (~938 °C), which prevents melt infiltration into carbon frameworks; sluggish redox kinetics; severe polysulfide dissolution; poor conductivity. Addressing these challenges requires integrated advances in binder design, electrode engineering, and cathode nanostructuring. The large first-charge overpotential due to the insulating LiOH/Li2O surface layer in Li2S-LSBs hinders activation and induces irreversible side reactions. Chapter 3 proposes mitigating the activation barrier by exploiting the reaction between polyvinylidene fluoride (PVDF) binder and LiOH/Li2O through dehydrofluorination. The overpotential was successfully reduced from 3.74 V with 30 min slurry grinding to 2.75 V by extending slurry stirring to 48 h. However, PVDF was also found to react with Li2S itself, partially consuming active material and lowering discharge capacity. Overall, this study provides mechanistic insights into the origin of Li2S activation overpotential and demonstrates the dual role of conventional PVDF binders, where slurry processing with PVDF can effectively reduce the first-charge barrier, while also highlighting the limitations of PVDF as a binder for Li2S electrodes. Since PVDF proved unsuitable for Li2S electrodes, Chapter 4 investigates alternative binders capable of enhancing the electrochemical performance of Li2S-LSBs. A binder based on a zinc acetate triethanolamine (Zn(OAc)2·TEA) complex was developed, which not only provides strong polysulfide-trapping ability but also exhibits redox catalytic activity, leading to markedly improved capacity, rate capability, and cycling stability compared with PVDF. To further reinforce electrode integrity and improve dispersion stability, polyethylenimine (PEI) was incorporated to form a Zn(OAc)2·TEA/PEI hybrid binder. Electrochemical testing showed that Li2S cathodes employing Zn(OAc)2·TEA/PEI with 10 wt.% PEI achieved superior rate performance, high discharge capacity, and excellent long-term cycling stability. An additional advantage of these binders is their fluorine-free composition, which aligns with sustainability goals and complying with emerging regulations, including EU restrictions on per- and polyfluoroalkyl substances (PFAS). In Chapter 5, an efficient precursor solution infiltration-decomposition strategy was invented to synthesize Li2S@Carbon nanocomposites under mild conditions, overcoming the challenges of Li2S’s high melting point, poor solubility, and the large particle size of commercial Li2S. In this approach, Li2S was first reacted with carbon disulfide (CS2) in ethanol at ambient temperature to form a highly soluble lithium trithiocarbonate (Li2CS3) precursor, which was readily infiltrated into mesoporous Super P carbon (SP). Subsequent thermal decomposition of Li2CS3@SP at 400 °C produced Li2S@SP-400 nanocomposites with a Li2S:SP mass ratio of 60:40, containing finely dispersed Li2S particles (~11 nm) uniformly confined within the Super P matrix. Electrochemical testing demonstrated that these nanocomposites delivered a high discharge capacity of 821 mA h g-1 (Li2S) at 0.1 C, equivalent to 1190 mA h g-1 (S), and exhibited superior rate capability and cycling stability compared to commercial Li2S, non-infiltrated Li2S nanoparticles, and melt-infiltrated sulfur composites (S@SP). The thermal decomposition of Li2CS3 precursor releases a large amount of CS2 gas (~62 wt.% of the precursor), which creates internal voids and limits the in-pore Li2S loading. To address this, Chapter 6 builds upon precursor infiltration-decomposition method with a multi-cycle strategy, enabling higher Li2S content and in-pore loading. Using mesoporous Super P as the conductive host and Li2CS3 as the precursor, repeated infiltration-decomposition cycles progressively increased the pore filling factor (FF) and in-pore Li2S loading (IPL), from FF = 38% and IPL = 30% for Li2S@SP-1 (one cycle) to FF = 91% and IPL = 73% for Li2S@SP-5 (five cycles), while also raising the overall Li2S content to 70 wt.%. Direct structural evidence from XRD and SEM confirmed reduced crystallite size, suppressed external deposition, and uniform Li2S distribution in the optimized Li2S@SP-5. Electrochemical tests demonstrated that Li2S@SP-5 delivered an initial discharge capacity of 807 mA h g-1 (Li2S) at 0.1 C, 598 mA h g-1 (Li2S) in the first cycle at 1.0 C, and retained 376 mA h g-1 (Li2S) after 500 cycles at 1.0 C. To construct high-performance cathodes, the functional binder from Chapter 4 was combined with the high in-pore loading Li2S@SP from Chapter 6. This attempt failed because Zn(OAc)2·TEA/PEI-based binders exhibited limitations with highly reactive nanoscale Li2S, resulting in diminished binding effectiveness. Chapter 7 therefore introduces a series of polyethylenimine-epoxy resin (PEI-ER) binders, where high-molecular-weight PEI anchors and catalyzes polysulfides while epoxy crosslinking reinforces mechanical stability, making this strategy particularly effective for stabilizing nanoscale Li2S composites. The in-situ crosslinking method further improved processing by removing the short crosslinking time window and enabling uniform networks without altering Li2S@SP morphology. Electrochemical tests showed the optimized in-situ crosslinked PEI-ER1:1 binder achieved 928 mA h g-1 at 0.05 C, 688 mA h g-1 in the first cycle at 0.5 C and retained 325 mA h g-1 after 1000 cycles at 0.5 C with stable Coulombic efficiency. SEM confirmed its compact structure, establishing in-situ PEI-ER crosslinking as a robust binder strategy for nanoscale, high-loading Li2S cathodes. Chapter 8 serves as the culmination of these research projects, combining the optimized Li2S@Carbon cathodes from Chapter 6 and functional binders developed from Chapter 7 with commercial Si/C anodes to successfully assemble and evaluate lithium-anode-free full cells, with PVP used as a baseline comparison, thereby demonstrating their practical feasibility. The in-situ crosslinked PEI-ER1:1-based full cell batteries delivered 670 mA h g-1 at 0.1 C and retained 304 mA h g-1 after 100 cycles (~45% retention), outperforming PVP-based full cell batteries (582 to 250 mA h g-1, ~43%). At 0.5 C, the in-situ crosslinked PEI-ER1:1-based full cell batteries achieved 564 mA h g-1 after activation and maintained 377 mA h g-1 after 500 cycles (66.8% retention), while the PVP counterparts fell from 573 to 176 mA h g-1 (30.7%). These results underscore the binder’s role in stabilizing cathodes and mark the successful assembly of lithium-free-anode Li2S full cells with commercial Si/C anodes. In summary, this thesis addresses the critical challenges of Li2S cathodes, including the large first-charge overpotential, the drawback of PVDF consuming Li2S, the large particle size of commercial Li2S, the high melting point and poor solubility that hinder conventional Li2S@Carbon composite fabrication, and the limitations of binders when applied to nanoscale Li2S, each identified in the process of resolving the preceding issue. By systematically investigating these problems, this thesis advances functional binder design, exploits precursor chemistry, and engineers nanostructured composites, concluding with the successful demonstration of lithium-anode-free full cell batteries. Further improvements could be achieved by employing more efficient carbon hosts with tailored structures, developing high-loading electrodes, integrating solid-state electrolytes to mitigate polysulfide dissolution, and incorporating catalytic components to accelerate Li2S redox kinetics, thereby pushing Li2S-LSBs closer to practical, high-energy-density applications.
Validating & Measuring Influenza Vaccine Effectiveness among and against Cardiovascular Hospitalization
(University of Waterloo, 2025-11-19) Amoud, Razan
Background: Influenza is a viral respiratory infection that causes serious health outcomes such as hospitalization and death and represents an important public health burden globally. Vaccination is one of the most effective interventions to prevent influenza and its complications. Although administrative databases such as pharmacy billing claims are used to measure influenza vaccination status, little is known of the validity of these databases in Ontario. Patients with cardiovascular disease (CVD) may have an altered immune response, despite being at a higher risk of influenza complications. It is not known if the Vaccine Effectiveness (VE) in this population is comparable to the general population. Further, there is a lack of research in Canada evaluating influenza vaccine effectiveness against cardiovascular outcomes, particularly using robust study designs such as the test-negative design. Three interrelated studies were carried out. First, a validation study was completed to examine the accuracy of the combination of Ontario’s administrative data from pharmacy and physician billing claims in identifying an individual’s vaccination status. The second study assessed the influenza VE against laboratory-confirmed influenza among older adults hospitalized with CVD conditions in Ontario and examined sex and age group as potential effect modifiers in the association between influenza vaccination and laboratory-confirmed influenza. The third study measured influenza VE against acute CVD outcomes using the Test-Negative Design (TND) for the first time among older adults in Ontario who were hospitalized within three days of their influenza testing. Methods: In the first study, I validated the combined physician and pharmacy billing claims within administrative databases using the linked reference standard of self-report data from the Canadian Community Health Survey (CCHS). This study estimated sensitivity, specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV), with 95% Confidence Intervals (CI) of the estimates. In the second study, I used the TND to measure influenza VE against laboratory-confirmed influenza. The analysis included both crude and adjusted estimates accounting for potential confounders (sex, age group, neighbourhood income quintile, rurality, number of outpatient visits, beta-blocker medication use, statin medication use, receiving home care services, and influenza testing relative to the peak month of the season, and year of influenza season). In addition, I assessed effect modification of sex and age group on influenza VE by using two different methods: introducing interaction terms into the model and stratification. In the third study, I examined influenza VE against acute CVD outcomes, also using the TND. The exposure of interest was vaccination status, cases were those hospitalized for myocardial infarction, unstable angina, or stroke and testing positive for influenza, and controls were those hospitalized for a non-cardiovascular event and testing negative for influenza. Both crude and adjusted VE were estimated. Sensitivity analyses were performed to test the robustness of the findings by varying inclusion criteria, outcome definitions, and statistical adjustments. Results: In the first study, the CCHS identified 43% as vaccinated across the two survey cycles of 2013 and 2014. The sensitivity for the combined pharmacy and physician billing codes was 60.1% (95% CI 59.3%–61.0%), specificity was 98.5% (95% CI 98.3%–98.7%), PPV was 96.7% (95% CI 96.3%–97.1%) and NPV was 76.9% (95% CI 76.4%–77.5%). The second study included 1,159 patients, and almost half were vaccinated. Among vaccinated patients, 14% tested positive for influenza compared to 20% of unvaccinated patients.. Crude and adjusted VE were 32% [95% CI, 8–50%] and 43% [95% CI 20%–60%], respectively. Neither inclusion of interaction terms separately in the full model, nor stratification by sex or age group, revealed any evidence of effect modification. The third study included 33,710 hospitalized individuals who tested positive for influenza and had a CVD event (cases) or tested negative and did not have a CVD event (controls). There were 18,519 vaccinated individuals (55%). Among vaccinated patients, 0.4% tested positive for influenza and were hospitalized for a cardiovascular event, compared to 0.8% of unvaccinated patients. The adjusted influenza VE against cardiovascular outcomes was 43% [95% CI 25%–58%; p-value =0.0001]. Conclusion: Compared to past studies with only physician billing claims, the validation study provided improved performance measures of sensitivity, specificity, PPV and NPV values in the combined physician and pharmacy billing claims in identifying individual vaccination status in Ontario. The second study estimated influenza VE against laboratory confirmed infection and supports that influenza VE among older adults with CVD hospitalization is comparable to the general population. Also, no significant effect modification in VE was observed by the patient’s sex, or age. These findings suggest that the protective effect of the influenza vaccine against laboratory-confirmed influenza is consistent across key demographic and clinical subgroups within this high-risk population. The third study found that the influenza vaccine provides a significant protective effect against CVD outcomes. The global findings of this thesis emphasize the validity of administrative databases at estimating population-level vaccination rates and show the importance of influenza vaccination as an effective strategy to reduce both influenza hospitalization and CVD events. This unique research equips healthcare providers and policy makers with relevant findings to support their campaigns and recommendations on influenza VE, particularly in relation to CVD outcomes.
Turing Instability of a Closed Nutrient-Phytoplankton-Zooplankton Model with Nutrient Recycling
(University of Waterloo, 2025-11-19) Xu, Xiangye
We investigate Turing instability in a closed Nutrient–Phytoplankton–Zooplankton (NPZ) ecosystem that incorporates delayed nutrient recycling, formulated as a reaction–diffusion system. Although spatial diffusion typically enhances system stability, our study focuses on how differing diffusion rates among species can destabilize steady states and lead to the emergence of spatial patterns. To explore this, we first perform a linear stability analysis to identify the conditions under which Turing instability arises. These theoretical predictions are then validated through numerical simulations. Our study progresses systematically: beginning with a two-species model, extending to a threespecies system, and finally to a four species NPZD model. This stepwise framework provides both conceptual insight and quantitative understanding of how diffusion influences instabilities, offering a comprehensive perspective on pattern formation in multi-species plankton ecosystems.
Mitigating Hardware Trojan Risks in the Global IC Supply Chain: Pre- and Post-Silicon Detection Approaches
(University of Waterloo, 2025-11-19) Pintur, Michael
The integrity of modern systems is critically dependent on trust in the underlying hardware, yet complex Integrated Circuit (IC) supply chains introduce numerous vulnerabilities for malicious insertions. This thesis confronts the challenge of IC trust by examining two distinct detection methodologies, illuminating the fundamental trade-offs inherent in practical hardware verification under black-box conditions. The first contribution targets Trojan detection in Third Party Intellectual Property (3PIP) by adapting power-based side-channel fuzzing with Field-Programmable Gate Arrays (FPGAs). This investigation confirms that dynamic power analysis serves as an effective oracle for identifying the activation of a Trojan, creating a statistically significant side-channel anomaly. However, the work also demonstrates that random fuzzing is an impractical search strategy for discovering the low-probability trigger required for activation, highlighting a significant barrier to its widespread adoption. To overcome the limitations of methods requiring dynamic Trojan activation, this work explores static, on-chip sensing using Ring Oscillator Networks (RONs). This research addresses a gap in prior work by characterizing RON behaviour on a modern 28nm process and subsequently developing a statistical framework to distinguish malicious modifications from normal process variations. The proposed approach was validated against a benchmark hardware Trojan and successfully classified all Trojan-free and Trojan-infected devices. These results confirm that RON-based detection remains effective on 28nm process technology and demonstrate the robustness of the developed anomaly detection algorithm. By juxtaposing a dynamic, trigger-based detection method with a static, reference-based approach, this thesis illuminates the fundamental trade-offs inherent in hardware trust verification. The findings reveal a practical difference between the high specificity of dynamic analysis and the broad applicability of static verification. This research concludes that while physical side-channels are powerful tools, future progress will depend on developing solutions that effectively balance these competing demands, for a more comprehensive security strategy in the IC supply chain.
Deep Learning and Dynamical Systems Approaches to Critical Transitions in Socio–Climate and Complex Systems
(University of Waterloo, 2025-11-19) Babazadeh Maghsoodlo, Yazdan
This thesis explores how dynamical systems, stochastic processes, and deep learning can be integrated to study critical transitions in socio-climate and other complex systems. Chapter 1 establishes the conceptual foundation, introducing complex systems, tipping points, bifurcation theory, stochasticity, early warning signals, and the role of deep learning. It also highlights flickering as a precursor to collapse and motivates the importance of coupled socio-climate feedbacks. Chapter 2 develops a hybrid CNN--LSTM framework to classify bifurcations in noisy time series. Trained on synthetic dynamical models, the classifier generalises to empirical data and outperforms traditional early warning signals, offering a robust method to identify fold, Hopf, and transcritical bifurcations. Chapter 3 introduces a deep learning approach to detect flickering dynamics, noise-driven switching between alternative equilibria. The model distinguishes true flickering from noise-induced variance inflation across diverse systems and demonstrates applicability to empirical data such as palaeoclimate records and physiological signals, providing an early warning beyond variance-based methods. Chapter 4 presents a coupled socio-climate model where social behaviour feeds back on emissions and climate thresholds. Results show that social dynamics, such as faster learning rates or stronger norms, can delay or prevent climate tipping, while delays or weak norms accelerate collapse. This chapter highlights the potential of social tipping points to stabilize climate trajectories. Chapter 5 evaluates whether binary opinion models suffice to represent socio-climate interactions compared to richer spectrum models. Using replicator and Friedkin–Johnsen frameworks coupled to climate-carbon and forest-grassland systems, the study finds that binary models capture essential coupled dynamics with surprising accuracy, despite their simplicity. Together, the chapters demonstrate that combining dynamical systems theory, stochastic analysis, and deep learning yields powerful tools to anticipate tipping points. The findings advance both methodological development and practical insight, showing that human social responses can critically shape whether climate transitions are mitigated or exacerbated.

Browse

Recent Submissions