Depth perception has become a heavily researched area as companies and researchers are striving towards the development of self-driving cars. Self-driving cars rely on perceiving the surrounding area, which heavily depends on technology capable of providing the system with depth perception capabilities. In this paper, we explore developing a single camera (monocular) depth prediction model that is trained on panoramic depth images. Our model makes novel use of transfer learning efficient encoder models, pre-training on a larger dataset of flat depth images, and optimizing the model for use with a Jetson Nano. Additionally, we present a training and optimization framework to make developing and testing new monocular depth perception models easier and faster. While the model failed to achieve a high frame rate, the framework and models developed are a promising starting place for future work.

Cyber attacks are a growing concern in our modern world, making security evaluation a critical venture. Penetration testing, the process of attempting to compromise a computer network with controlled tests, is a proven method of evaluating a system's security measures. However, penetration tests, and preventive security analysis in general, require considerable investments in money, time, and labor, which can cause them to be overlooked. Alternatively, automated penetration testing programs are used to conduct a security evaluation with less user effort, lower cost, and in a shorter period of time than manual penetration tests. The trade-off is that automated penetration testing tools are not as effective as manual tests. They are not as flexible as manual testing, cannot discover every vulnerability, and can lead to a false sense of security. The development of better automated tools can help organizations quickly and frequently know the state of their security measures and can help improve the manual penetration testing process by accelerating repetitive tasks without sacrificing results. This thesis presents Automated Network Exploitation through Penetration Testing (ANEX), an automated penetration testing system designed to infiltrate a computer network and map paths from a compromised network machine to a specified target machine. Our goal is to provide an effective security evaluation solution with minimal user involvement that is easily deployable in an existing system. ANEX demonstrates that important security information can be gathered through automated tools based solely on free-to-use programs. ANEX can also enhance the manual penetration testing process by quickly accumulating information about each machine to develop more focused testing procedures. Our results show that we are able to successfully infiltrate multiple network levels and exploit machines not directly accessible to our testing machine with mixed success. Overall, our design shows the efficacy of utilizing automated and open-source tools for penetration testing.

Coercivity is a characteristic property of the bilinear term in a weak form of a partial differential equation in both infinite space and the corresponding finite space utilized by a numerical scheme. This concept implies \textit{stability} and \textit{well-posedness} of the weak form in both the exact solution and the numerical solution. In fact, the loss of this property especially in finite dimension cases leads to instability of the numerical scheme. This phenomenon occurs in three major families of problems consisting of advection-diffusion equation with dominant advection term, elastic analysis of very thin beams, and associated plasticity and non-associated plasticity problems. There are two main paths to overcome the loss of coercivity, first manipulating and stabilizing a weak form to ensure that the discrete weak form is coercive, second using an automatically stable method to estimate the solution space such as the Discontinuous Petrov Galerkin (DPG) method in which the optimal test space is attained during the design of the method in such a way that the scheme keeps the coercivity inherently. In this dissertation, A stable numerical method for the aforementioned problems is proposed. A stabilized finite element method for the problem of migration risk problem which belongs to the family of the advection-diffusion problems is designed and thoroughly analyzed. Moreover, DPG method is exploited for a wide range of valuing option problems under the black-Scholes model including vanilla options, American options, Asian options, double knock barrier options where they all belong to family of advection-diffusion problem, and elastic analysis of Timoshenko beam theory. Besides, The problem of American option pricing, migration risk, and plasticity problems can be categorized as a free boundary value problem which has their extra complexity, and optimization theory and variational inequality are the main tools to study these families of the problems. Thus, an overview of the classic definition of variational inequalities and different tools and methods to study analytically and numerically this family of problems is provided and a novel adjoint sensitivity analysis of variational inequalities is proposed.

Over the past two decades there has been a rapid decline in public oversight of state and local governments. From 2003 to 2014, the number of journalists assigned to cover the proceedings in state houses has declined by more than 30\%. During the same time period, non-profit projects such as Digital Democracy sought to collect and store legislative bill and hearing information on behalf of the public. More recently, AI4Reporters, an offshoot of Digital Democracy, seeks to actively summarize interesting legislative data. This thesis presents STRAINER, a parallel project with AI4Reporters, as an active data retrieval and filtering system for surfacing newsworthy legislative data. Within STRAINER we define and implement a process pipeline by which information regarding legislative bill discussion events can be collected from a variety of sources and aggregated into feature sets suitable for machine learning. Utilizing two independent labeling techniques we trained a variety of SVM and Logistic Regression models to predict the newsworthiness of bill discussions that took place in the California State Legislature during the 2017-2018 session year. We found that our models were able to correctly retrieve more than 80\% of newsworthy discussions.

Collaboration has become a key feature of modern software, allowing teams to work together effectively in real-time while in different locations. In order for a user to communicate their intention to several distributed peers, computing devices must exchange high-frequency updates with transient metadata like mouse position, text range highlights, and temporary comments. Current peer-to-peer awareness solutions have high time and space complexity due to the ever-expanding logs that each client must maintain in order to ensure robust collaboration in eventually consistent environments. This paper proposes an awareness Conflict-Free Replicated Data Type (CRDT) library that provides the tooling to support an eventually consistent, decentralized, and robust multi-user collaborative environment. Our library is tuned for rapid iterative updates that communicate fine-grained user actions across a network of collaborators. Our approach holds memory constant for subsequent writes to an existing key on a shared resource and completely prunes stale data from shared documents. These features allow us to keep the CRDT's memory footprint small, making it a feasible solution for memory constrained applications. Results show that our CRDT implementation is comparable to or exceeds the performance of similar data structures in high-frequency read/write scenarios.

<p>Parallel MRI, in which k-space is regularly or irregularly undersampled, is critical for imaging speed acceleration. In this thesis, we show how to optimize a regular undersampling pattern for three-dimensional Cartesian imaging in order to achieve faster data acquisition and/or higher signal to noise ratio (SNR) by using nonlinear optimization. A new sensitivity profiling approach is proposed to produce better sensitivity maps, required for the sampling optimization. This design approach is easily adapted to calculate sensitivities for arbitrary planes and volumes. The use of a semi-definite, linearly constrained model to optimize a parallel MRI undersampling pattern is novel. To solve this problem, an iterative trust-region is applied. When tested on real coil data, the optimal solution presents a significant theoretical improvement in accelerating data acquisition speed and eliminating noise.</p> / Master of Applied Science (MASc)

Ambient temperature sensing plays an important role in a number of applications in agriculture, industry, daily health care. In this thesis project, we propose a new acoustic-based ambient temperature sensing method called Mtemp. Mtemp empowers acoustic-enabled IoT devices, smartphones to perform ambient air temperature sensing without additional hardware. Basically, Mtemp utilizes on-board speaker and microphone to calculate the propagation speed of acoustic signal by measuring the phrase of the target signal, thereby estimate the ambient temperature according to a roughly linear relationship between temperature and sound speed. Mtemp is portable and economical, making it competitive compared with traditional thermometers for ubiquitous sensing.

This thesis reports a comprehensive study of quaternary BeMgZnO alloy and BeMgZnO/ZnO heterostructures for UV-optoelectronics electronic applications. It was shown that by tuning Be and Mg contents in the heterostructures, high carrier densities of two-dimensional electron gas (2DEG) are achievable and makes its use possible for high power RF applications. Additionally, optical bandgaps as high as 5.1 eV were achieved for single crystal wurtzite material which allows the use of the alloy for solar blind optoelectronics (Eg>4.5eV) or intersubband devices. A systematic experimental and theoretical study of lattice parameters and bandgaps of quaternary BeMgZnO alloy was performed for the whole range of compositions. Composition independent bowing parameters were determined which allows accurate predictions of experimentally measured values. The BeMgZnO thin films were grown by plasma assisted molecular beam epitaxy (P-MBE) in a wide range of compositions. The optimization of the growth conditions and its effects on the material properties were explored. The surface morphology and electrical characteristics of the films grown on (0001) sapphire were found to critically depend on the metal-to-oxygen ratio. Samples grown under slightly oxygen-rich conditions exhibited the lowest RMS surface roughness (as low as 0.5 nm). Additionally, the films grown under oxygen-rich conditions were semi-insulating (>105 Ω∙cm), while the films grown under metal-rich conditions were semiconducting (~102 Ω∙cm). Additionally, with increasing bandgap Stokes shift increases, reaching ~0.5 eV for the films with 4.6 eV absorption edge suggests the presence of band tail states introduced by potential fluctuations and alloying. From spectrally resolved PL transients, BeMgZnO films grown on a GaN/sapphire template having higher Mg/Be content ratio exhibit smaller localization depth and brighter photoluminescence at low temperatures. The optimum content ratio for better room temperature optical performance was found to be ~2.5. The BeMgZnO material system and heterostructures are promising candidates for the device fabrication. 2DEG densities of MgZnO/ZnO heterostructures were shown to improve significantly (above 1013 cm-2) by adding even a small amount of Be (1-5%). As an essential step toward device fabrication, reliable ohmic contacts to ZnO were established with remarkably low specific contact resistivities below 10-6 Ohm-cm2 for films with 1018 cm-3 carrier density.

In the modern advanced gas turbines, the turbine inlet temperature may exceed 1500°C as a requirement to increase power output and thermal efficiency. Therefore, it is imperative that the blades and vanes are cooled so they can withstand these extreme temperatures. Film cooling is a cooling technique widely used in high-performance gas turbines. However, the film cooling effectiveness has almost reached plateau, resulting in a bottleneck for continuous improvement of gas turbines' efficiency. In this study, an innovative cooling scheme, mist film cooling is investigated through experiments. A small amount of tiny water droplets with an average diameter about 10-15 µm (mist) is injected into the cooling air to enhance the cooling performance. A Phase Doppler Particle Analyzer (PDPA) system is used for droplet measurements. Mist film cooling performance is evaluated and compared against air-only film cooling. This study continues the previous work by (a) adding fan-shaped holes and comparing their cooling performance with the round holes, (b) extending the length of the test section to study the performance farther downstream the injection holds, and (c) using computational simulation to investigate the feasibility of transporting mist to the film cooling holes through gas turbine inside passages. The results show that, with an appropriate blowing ratio, the fan-shaped holes performs about 200% better than round holes in cooling effectiveness and adding 10% (wt.) mist can further enhance cooling effectiveness 170% in average. Farther downstream away from the injection holes (X/D> 50), mist cooling enhancement prevails and actually increases significantly. PDPA measurements have shed lights to the fundamental physics of droplet dynamics and their interactions with thermo-flow fields. These experimental results lead to either using lower amount of cooling air or use fewer number of cooing holes rows. This means higher gas turbine power output, higher thermal efficiency, and longer components life which will reflect as a cheaper electricity bill. Computational Fluid Dynamics (CFD) showed that it is feasible to transport the water mist, with initial diameters ranging from 30 µm-50 µm and mist ratio of 10-15%, to the cooling holes on the surface of the turbine vanes and rotors to provide the desired film cooling. Key words: Gas Turbines, Heat Transfer, Film / mist Cooling, Experimental Study, Mist Transport, CFD, PDPA.

The application of Computer Simulation and Animation (CSA) in the instruction of engineering dynamics has shown a significant growth in the recent years. The two foremost methods to evaluate the effectiveness of CSA tools, including student feedback and surveys and measuring student change in performance, suggest that CSA modules improve student learning in engineering dynamics. However, neither method fully demonstrates the quality of students’ cognitive changes. This study examined the quality of effects of application of CSA modules on student learning and problem solving in particle dynamics. It also compared CSA modules with textbook-style problem-solving regarding the changes they cause in students’ cognitive process. A qualitative methodology was adopted to design and implement a study to explore the changes in participants’ learning and problem-solving behavior caused by using a CSA module. Collected data were coded and analyzed using the categories of cognitive process based on the Revised Bloom’s Taxonomy. An analysis of the results revealed that the most significant effects were observed in understanding, analyzing, and evaluating. The high frequency of “inference” behavior after working with modules indicated a significant increase in participants’ understanding activity after working with computer modules. Comparing behavior changes of computer-simulation group students with those who worked with a textbook-style example demonstrated that the CSA modules ignited more analytical behavior among students than did textbook-style examples. This study illustrated that improvement in learning due to the application of CSA is not limited to conceptual understanding; CSA modules enhance students’ skills in applying, organizing, and evaluating as well. The interactive characteristics of CSA play a major role in stimulating students’ analytical reasoning and critical thinking in engineering dynamics.