Carleton University Institutional Repository
Not a member yet
20816 research outputs found
Sort by
Electro-immobilisation, chemical anaesthesia, and the complex future of wild fish welfare in research
Get PDFIn research, fish must undergo invasive procedures or otherwise extensive handling for which some form of anaesthesia or immobilisation is typically required. Chemical anaesthetics are commonly used for such purposes, and electrical methods of fish immobilisation (“electro-immobilisation”) have received attention as potential alternatives to chemical anaesthesia in many settings. Both chemical anaesthetics and electro-immobilisation techniques are associated with benefits and challenges with respect to different facets of fish welfare, logistics in field scenarios, and legal status. The ultimate purpose of this thesis was to examine some of the overlooked or understudied considerations (both biological and technological) that are relevant to formulating evidence-based assessments of best anaesthesia and immobilisation practices in wild fish research, and attempt to generate such an assessment by uniting some of the key lessons and challenges gleaned from studying fish welfare in this context. After a general introduction to the topic, I describe two experiments detailing novel comparisons of chemical anaesthesia and electro-immobilisation methods. The first focused on cardiac activity in walleye (Sander vitreus) measured continuously over the course of anaesthetic/immobilisation induction and recovery periods, and the second used direct live scoring to assess impairment of parental care behaviours and nest abandonment in nesting smallmouth bass (Micropterus dolomieu) following different anaesthetic or immobilisation treatments. I then recount an experiment evaluating the efficacy and shortcomings associated with a custom-built electro-immobilisation apparatus tested on largemouth bass (M. nigricans), with implications for practical applications and apparatus designs. Next, I present an experimental comparison of several chemical anaesthetics and immobilisation methods in a non-teleost fish, juvenile lake sturgeon (Acipenser fulvescens), that entailed quantifying physiological stress and behavioural alterations following standardised surgeries under focal methods. In the penultimate chapter, empirical research, ethics, and contemporary debates in fish welfare combine in the form of a novel characterisation of the precautionary principle and an anaesthetic/immobilisation method selection framework designed to provide a potential path forward for assessing “optimal” approaches to anaesthesia and electro-immobilisation in wild fishes on a case-by-case basis. I conclude with a summary of the works detailed herein, future research recommendations, and some ruminations on the future of electro-immobilisation in fish research
Comparing Parameter Estimation and State Prediction Performance of Physics Informed Neural Networks in Relation to Bayesian Inference
Get PDFThis thesis investigates two physics informed neural network (PINN) approaches for two mechanistic problems via the performance metrics of model parameter estimation and system state prediction. The PINNs are compared to Bayesian inference (BI) tested under similar conditions. The first PINN studied is the loss-based PINN, which embeds physics into a standard feedforward neural network within the loss function of this network. The second PINN investigated is the hybrid-recurrent physics informed neural network, which uses a recurrent neural network (RNN) that embeds physics in RNN architecture. The BI used as a point of comparison uses a Transitional Markov Chain Monte Carlo algorithm to estimate the posterior distributions. The applications studied are a compartmental model for an infectious disease, and a mass-spring-damper model. It is found that BI performs better in ideal data conditions and the PINNs perform better in weak data conditions
Nonlinear Disturbance Estimation-Enhanced Simple Adaptive Control of a Space Robot
No full textThis thesis proposes a control scheme for a space robot system combining a simple adaptive controller (SAC) with a non-linear disturbance observer (NDO) to facilitate end-effector trajectory tracking. The proposed control scheme is implemented to facilitate spacecraft capture, demonstrating its utility for debris removal. The study begins by defining a six-degree-of-freedom planar free-floating space robot system through its Euler-Lagrange kinematic and dynamic models. The SAC control law is developed with computationally efficient modifications while ensuring optimal performance. Furthermore, the NDO integrates theories for rigid body and manipulator systems to compensate for disturbances, including joint friction and unknown external disturbances. Simulations are conducted to validate the performance of the NDO-enhanced SAC scheme in various test cases
The Geomorphology and Provenance of the Exeter Lake Esker
No full textA system-scale examination of the geomorphology and provenance of the Exeter Lake Esker (ELE) was performed using a large integrated dataset to better understand how eskers form. The ELE consists of three landform types: Type 1 ridges – narrow, pebbly to bouldery rounded to sharp-crested ridges, (2) Type 2 ridges – flat- to irregular-topped ridges that are wider and finer grained (sandy to cobbly), and (3) Type 3 associated sediment – finer grained (sandy to pebbly) enlargements and ridge-attached and detached broader sediment bodies. The transition from Type 1 to Type 2 and 3 landforms occurs over distances of ~ 5 – 20 km. Stratigraphically, Type 1 landforms are overlain by Type 2 and Type 3 deposits. Systematic, system-scale changes in ridge size, and grain size and shape, were not observed. The provenance of the ELE was examined using 54 paired esker and till samples collected along the length of the esker. Several dispersal trains are observed in both till and esker, the most well-defined of which emanates from the Thelon Formation, an SiO2 rich siliciclastic unit. The concentration of Thelon Formation pebbles in both the esker and till samples decreases rapidly from 100% to <50% ~25 km down-ice from the bedrock source and approaches 0% 100 km down-ice from the bedrock source. Likewise, the concentration of SiO2 in the mud (<0.063 mm) fraction in both esker and till decreases rapidly by ~25 km down-ice from the bedrock source and levels out to background 125 km down-ice of the bedrock source. Given the weight of evidence—the repeated transitions from Type 1 (proximal) to Type 2 (medial) and Type 3 (distal) deposits, the lack of system scale down-esker increases in esker size or grain rounding, the lack of a system-scale down-esker fining, the mirroring of esker and till geochemical values over 750 km, the general absence of dispersal train extension in the esker relative to the till, and the general absence dispersal train extension between the pebble and mud size fractions in the esker—the esker system was most likely deposited in short segments 5 to 25 km in length
A deep learning pipeline for real-time palliative radiotherapy of spine metastasis
No full textThe standard palliative radiotherapy workflow follows a scan-plan-treat approach relying on dedicated CT simulation scans, introducing delays and inconvenience, particularly for patients with spinal metastases. Efforts to achieve real-time palliative therapy have been limited to simpler treatment plans, or require eligible prior diagnostic imaging. Onboard cone-beam CT (CBCT) enables real-time treatment planning; however, higher conformality treatments require a rapid method for image quality improvement and relevant structure segmentation. This work leverages deep learning models to develop and evaluate a pipeline for synthetic CT (sCT) generation and automated vertebral segmentation. The pipeline begins with image quality enhancements using a conditional generative adversarial network (GAN)-based network for reducing streaking artifacts and a cycle-consistent GAN for CBCT-based sCT generation. Subsequent phases involve a three-stage network for vertebral segmentation, using heatmap regression techniques for coarse localization and fine identification of vertebrae, followed by segmentation using a U-Net architecture. The sCT generation phase achieved significant improvements in error metrics, signal-to-noise ratio, and structural similarity over CBCT. Dosimetric analysis revealed reduced dose discrepancies and improved gamma pass rates. The segmentation network, trained and validated on the VerSe Challenge datasets, demonstrated high identification rates and accurate segmentation metrics and was applied to sCT images. The ability to generate sCT images and perform accurate segmentation in real-time marks a significant advancement toward more conformal, real-time palliative therapy for spinal metastases. This method eliminates reliance on prior CT scans and simplifies the treatment process, offering a promising solution for quicker, more accessible palliative care
Sampling lessons learned from surveying by one Self-governing Indigenous Government (SGIG), a pilot project by the Nisga’a Nation
No full textAnalysis of Geography-Oriented Internet Routing
Get PDFThe Border Gateway Protocol (BGP) routing policy often leads to “boomerang routing” where traffic unnecessarily crosses jurisdictional boundaries, even when the source and destination are geographically close. This routing behavior raises privacy concerns due to exposure to varying data sovereignty laws. This thesis explores geography-oriented routing as a strategy to minimize privacy risks. It investigates the costs of avoiding specific geographic regions and the impact on performance when boomerang routing is replaced with more direct, boomerang-free routes. A city map, constructed using RIPE Atlas data, is used to assess metrics such as latency and city hop-count. The findings reveal that avoiding certain jurisdictions improves privacy but increases latency and city hop counts. Replacing boomerang routing with alternatives enhances privacy at a performance cost. This trade-off between privacy and routing feasibility underscores the need for further research to develop strategies that balance privacy concerns with performance requirements
The Accountants of Livent: Newspaper Representations of Accountants in a High Visibility Financial Scandal
No full textThis research investigates newspaper representations of six accountants implicated in the Livent Inc. (Livent) fraud scandal and named in charges laid by the United States Securities and Exchange Commission (SEC) in 1999. Newspaper coverage of these accountants presents a unique research opportunity because all six individuals share a common context in which they provided testimony: the 2008 criminal trial of the company’s founders. This group of accountants possessed common and contrasting individual characteristics, qualifications, and perspectives. Drawing on the sociology of professions literature, theories about how legitimacy judgments are made, and media theory, this research investigates whether, and in what ways, identifiable differences exist in how newspaper reports represented these individuals. Using a critical discourse analytical approach to focus on how the individual accountants were named (nomination) and described (predication) in newspaper articles, several themes emerged. Notably, gender and status appear as important distinctions that underlie some of the differences in representation made apparent through comparative analysis. Differences in representation are particularly evident around individual credibility. Taken together, these differences suggest that the newspapers do not name or describe these six individuals as representatives of the accounting profession. Rather the newspaper representations focused on their individual positions, characters, and gender. This study contributes a different approach through which the accounting profession can access public opinion and public perception and illuminates implications for accounting’s professional identity and professional legitimacy
Orientation Design to Support the Experiences of People Living with Dementia in Respite Houses
No full textOrientation and wayfinding are fundamental needs for people living with dementia, especially in new, unfamiliar environments such as Respite Houses. This qualitative case study explores orientation design to support people living with dementia in a Respite House located in Ottawa, Ontario, Canada through a holistic lens. Phase one involved gathering insights on environmental design and orientation considerations for accommodations serving people with dementia, which resulted in a theoretical framework to guide subsequent phases. In phase two, data was collected from staff members and clients of the Respite House through naturalistic observation and semi-structured interviews. Phase three involved co-design sessions to help inform an interior design that might enhance orientation in the Respite House at Perley Health. The study identifies specific needs of the Respite House clients and staff shaping a proposed design plan with practical short-term recommendations. Long-term recommendations are also outlined as areas for future research