Matching Items (1,645)
Filtering by
- All Subjects: Electrical Engineering
- All Subjects: Education
- Member of: ASU Electronic Theses and Dissertations
Description
As higher education embraces innovative educational models, support for the faculty members who must carry them out remains a vital ingredient for success. Despite this need, many institutions adopt innovations such as blended learning for all of the benefits afforded, with minimal consideration to meaningfully equip professors teaching these courses. “Faculty Learning Communities” (FLC’s) provide a powerful model of supporting and equipping faculty in their teaching practice. Nevertheless, ongoing and collaborative faculty development was historically unavailable to professors teaching undergraduate blended courses at Lancaster Bible College. Thus, the purpose of this qualitative action research study was to examine the ways that faculty perceived an FLC during the design and facilitation of a blended course. The Community of Inquiry (CoI) framework guided the design and facilitation of the FLC in fall 2018, as well as providing insight into measuring how learning communities formed during the FLC and while participants taught their courses. This FLC model blended learning for participants by occurring four times on campus, with online sessions following each in-person meeting. The faculty developer provided resources and support as faculty collaborated in designing their blended courses for the spring 2019 semester. Faculty perceptions of support were gathered in a focus group at the end of fall semester. During the spring 2019 semester, the faculty developer observed both on-campus and online sessions of the blended courses and led a second focus group about faculty perceptions of effectiveness and support. Qualitative data sets included video recordings of the FLC, focus groups, and class observations, field notes, and screenshots of online environments during the FLC and courses. Findings demonstrated substantial evidence of CoI measures of social presence, cognitive presence, and teaching presence were present in both the FLC and participants’ courses. These results affirmed the CoI framework provided a meaningful platform for faculty development. Additionally, participants perceived the FLC as supportive for their blended teaching practices, making direct mentions of support and indicating belief that broader institutional change be implemented toward this end to enhance faculty development opportunities. Limitations and implications of the study, as well as desired future research were explored.
ContributorsHarbin, Justin (Author) / Rotheram-Fuller, Erin (Thesis advisor) / Foulger, Teresa (Committee member) / Clawson, Penny (Committee member) / Buss, Ray R (Committee member) / Arizona State University (Publisher)
Created2019
Description
Chi and Wylie’s (2014) Interactive Constructive Active Passive Framework (ICAP) was used as the foundation of a teacher led intervention using small group instruction with manipulatives during mathematics instruction to provide developmentally appropriate instruction to kindergarten students in a rigorous academic program. This action research mixed-methods study was conducted in a full-day self-contained kindergarten classroom to ascertain the effects of this mathematics instruction method on students’ levels of engagement and attitudes. Over the course of six months, twenty mathematics lessons were recorded to gather data for the study. Quantitative data included measuring time-on-task, teacher behaviors ICAP level, student behaviors ICAP level, as well as a Student Attitude Survey that was conducted at the conclusion of the study. The Student Attitude Survey was presented in a modified Likert Scale format due to the age and reading ability of the participants. Qualitative data was gathered in the form of lesson transcripts. Twenty-two students and one classroom teacher participated in the study. Students ranged in age from five to six years old, and eleven participants (50%) were male. The results of the study showed that the use of small group hands-on instruction in mathematics had a positive effect on student engagement based on students’ time-on-task during the activity, as well as positive student attitudes toward mathematics as indicated on the Student Attitude Survey. Lesson transcripts and both teacher and student ICAP rubrics provided further support for the innovation.
ContributorsMiller, Jessica Ann (Author) / Rotheram-Fuller, Erin (Thesis advisor) / Oakes, Wendy (Committee member) / Wheeler, Melanie (Committee member) / Arizona State University (Publisher)
Created2019
Description
Zero-Valent Metals (ZVM) are highly reactive materials and have been proved to be effective in contaminant reduction in soils and groundwater remediation. In fact, zero-Valent Iron (ZVI) has proven to be very effective in removing, particularly chlorinated organics, heavy metals, and odorous sulfides. Addition of ZVI has also been proved in enhancing the methane gas generation in anaerobic digestion of activated sludge. However, no studies have been conducted regarding the effect of ZVM stimulation to Municipal Solid Waste (MSW) degradation. Therefore, a collaborative study was developed to manipulate microbial activity in the landfill bioreactors to favor methane production by adding ZVMs. This study focuses on evaluating the effects of added ZVM on the leachate generated from replicated lab scale landfill bioreactors. The specific objective was to investigate the effects of ZVMs addition on the organic and inorganic pollutants in leachate. The hypothesis here evaluated was that adding ZVM including ZVI and Zero Valent Manganese (ZVMn) will enhance the removal rates of the organic pollutants present in the leachate, likely by a putative higher rate of microbial metabolism. Test with six (4.23 gallons) bioreactors assembled with MSW collected from the Salt River Landfill and Southwest Regional Landfill showed that under 5 grams /liter of ZVI and 0.625 grams/liter of ZVMn additions, no significant difference was observed in the pH and temperature data of the leachate generated from these reactors. The conductivity data suggested the steady rise across all reactors over the period of time. The removal efficiency of sCOD was highest (27.112 mg/lit/day) for the reactors added with ZVMn at the end of 150 days for bottom layer, however the removal rate was highest (16.955 mg/lit/day) for ZVI after the end of 150 days of the middle layer. Similar trends in the results was observed in TC analysis. HPLC study indicated the dominance of the concentration of heptanoate and isovalerate were leachate generated from the bottom layer across all reactors. Heptanoate continued to dominate in the ZVMn added leachate even after middle layer injection. IC analysis concluded the chloride was dominant in the leachate generated from all the reactors and there was a steady increase in the chloride content over the period of time. Along with chloride, fluoride, bromide, nitrate, nitrite, phosphate and sulfate were also detected in considerable concentrations. In the summary, the addition of the zero valent metals has proved to be efficient in removal of the organics present in the leachate.
ContributorsPandit, Gandhar Abhay (Author) / Cadillo – Quiroz, Hinsby (Thesis advisor) / Olson, Larry (Thesis advisor) / Boyer, Treavor (Committee member) / Arizona State University (Publisher)
Created2019
Description
This dissertation examines the first impressions that occur between Deaf consumers and American Sign Language (ASL)/English interpreters prior to a healthcare appointment. Negative first impressions can lead to a disconnect or loss of trust between Deaf consumers and interpreters and increase the risk for Deaf consumers to receive inadequate healthcare. The recognition of this risk led to an action research study to look at barriers to successful interactions between ASL/English interpreters and Deaf consumers. The mixed methods research design and associated research questions discovered factors and perceptions that contributed to the disconnect and subsequently informed a 10-week intervention with a small group of ASL/English interpreters and Deaf consumers. The factors that influence connection are system related and a lack of a standardized approach to using name badges, missing or incorrect appointment details, and an inconsistent protocol for interpreter behavior when a healthcare provider leaves the room. The intervention allowed the interpreter participants to generate solutions to mitigate these barriers to connection and apply them during the 10 weeks. Deaf consumer feedback was gathered during the intervention period and was used to modify the generated solutions. The generated solutions included re-design of an interpreter referral agency’s name badge, using small talk as a way to learn information about the nature of the healthcare appointment and proactively discuss procedures when a healthcare provider leaves the exam room. These solutions resulted in a positive influence for both interpreters and Deaf consumers and an increase of trust and connection. The findings of this study show new approaches that create a connection between interpreters and Deaf consumers and may lead to more satisfactory healthcare interactions for Deaf consumers.
ContributorsCovey von Pingel, Teddi Lynn (Author) / Bertrand, Melanie (Thesis advisor) / Bernstein, Katie A (Committee member) / Roberson, Len (Committee member) / Arizona State University (Publisher)
Created2019
Description
This thesis presents efficient implementations of several linear algebra kernels, machine learning kernels and a neural network based recommender systems engine onto a massively parallel reconfigurable architecture, Transformer. The linear algebra kernels include Triangular Matrix Solver (TRSM), LU Decomposition (LUD), QR Decomposition (QRD), and Matrix Inversion. The machine learning kernels include an LSTM (Long Short Term Memory) cell, and a GRU (gated Recurrent Unit) cell used in recurrent neural networks. The neural network based recommender systems engine consists of multiple kernels including fully connected layers, embedding layer, 1-D batchnorm, Adam optimizer, etc.
Transformer is a massively parallel reconfigurable multicore architecture designed at the University of Michigan. The Transformer configuration considered here is 4 tiles and 16 General Processing Elements (GPEs) per tile. It supports a two level cache hierarchy where the L1 and L2 caches can operate in shared (S) or private (P) modes. The architecture was modeled using Gem5 and cycle accurate simulations were done to evaluate the performance in terms of execution times, giga-operations per second per Watt (GOPS/W), and giga-floating-point-operations per second per Watt (GFLOPS/W).
This thesis shows that for linear algebra kernels, each kernel achieves high performance for a certain cache mode and that this cache mode can change when the matrix size changes. For instance, for smaller matrix sizes, L1P, L2P cache mode is best for TRSM, while L1S, L2S is the best cache mode for LUD, and L1P, L2S is the best for QRD. For each kernel, the optimal cache mode changes when the matrix size is increased. For instance, for TRSM, the L1P, L2P cache mode is best for smaller matrix sizes ($N=64, 128, 256, 512$) and it changes to L1S, L2P for larger matrix sizes ($N=1024$). For machine learning kernels, L1P, L2P is the best cache mode for all network parameter sizes.
Gem5 simulations show that the peak performance for TRSM, LUD, QRD and Matrix Inverse in the 14nm node is 97.5, 59.4, 133.0 and 83.05 GFLOPS/W, respectively. For LSTM and GRU, the peak performance is 44.06 and 69.3 GFLOPS/W.
The neural network based recommender system was implemented in L1S, L2S cache mode. It includes a forward pass and a backward pass and is significantly more complex in terms of both computational complexity and data movement. The most computationally intensive block is the fully connected layer followed by Adam optimizer. The overall performance of the recommender systems engine is 54.55 GFLOPS/W and 169.12 GOPS/W.
Transformer is a massively parallel reconfigurable multicore architecture designed at the University of Michigan. The Transformer configuration considered here is 4 tiles and 16 General Processing Elements (GPEs) per tile. It supports a two level cache hierarchy where the L1 and L2 caches can operate in shared (S) or private (P) modes. The architecture was modeled using Gem5 and cycle accurate simulations were done to evaluate the performance in terms of execution times, giga-operations per second per Watt (GOPS/W), and giga-floating-point-operations per second per Watt (GFLOPS/W).
This thesis shows that for linear algebra kernels, each kernel achieves high performance for a certain cache mode and that this cache mode can change when the matrix size changes. For instance, for smaller matrix sizes, L1P, L2P cache mode is best for TRSM, while L1S, L2S is the best cache mode for LUD, and L1P, L2S is the best for QRD. For each kernel, the optimal cache mode changes when the matrix size is increased. For instance, for TRSM, the L1P, L2P cache mode is best for smaller matrix sizes ($N=64, 128, 256, 512$) and it changes to L1S, L2P for larger matrix sizes ($N=1024$). For machine learning kernels, L1P, L2P is the best cache mode for all network parameter sizes.
Gem5 simulations show that the peak performance for TRSM, LUD, QRD and Matrix Inverse in the 14nm node is 97.5, 59.4, 133.0 and 83.05 GFLOPS/W, respectively. For LSTM and GRU, the peak performance is 44.06 and 69.3 GFLOPS/W.
The neural network based recommender system was implemented in L1S, L2S cache mode. It includes a forward pass and a backward pass and is significantly more complex in terms of both computational complexity and data movement. The most computationally intensive block is the fully connected layer followed by Adam optimizer. The overall performance of the recommender systems engine is 54.55 GFLOPS/W and 169.12 GOPS/W.
ContributorsSoorishetty, Anuraag (Author) / Chakrabarti, Chaitali (Thesis advisor) / Kim, Hun Seok (Committee member) / LiKamWa, Robert (Committee member) / Arizona State University (Publisher)
Created2019
Description
Power management integrated circuit (PMIC) design is a key module in almost all electronics around us such as Phones, Tablets, Computers, Laptop, Electric vehicles, etc. The on-chip loads such as microprocessors cores, memories, Analog/RF, etc. requires multiple supply voltage domains. Providing these supply voltages from off-chip voltage regulators will increase the overall system cost and limits the performance due to the board and package parasitics. Therefore, an on-chip fully integrated voltage regulator (FIVR) is required.
The dissertation presents a topology for a fully integrated power stage in a DC-DC buck converter achieving a high-power density and a time-domain hysteresis based highly integrated buck converter. A multi-phase time-domain comparator is proposed in this work for implementing the hysteresis control, thereby achieving a process scaling friendly highly digital design. A higher-order LC notch filter along with a flying capacitor which couples the input and output voltage ripple is implemented. The power stage operates at 500 MHz and can deliver a maximum power of 1.0 W and load current of 1.67 A, while occupying 1.21 mm2 active die area. Thus achieving a power density of 0.867 W/mm2 and current density of 1.377 A/mm2. The peak efficiency obtained is 71% at 780 mA of load current. The power stage with the additional off-chip LC is utilized to design a highly integrated current mode hysteretic buck converter operating at 180 MHz. It achieves 20 ns of settling and 2-5 ns of rise/fall time for reference tracking.
The second part of the dissertation discusses an integrated low voltage switched-capacitor based power sensor, to measure the output power of a DC-DC boost converter. This approach results in a lower complexity, area, power consumption, and a lower component count for the overall PV MPPT system. Designed in a 180 nm CMOS process, the circuit can operate with a supply voltage of 1.8 V. It achieves a power sense accuracy of 7.6%, occupies a die area of 0.0519 mm2, and consumes 0.748 mW of power.
The dissertation presents a topology for a fully integrated power stage in a DC-DC buck converter achieving a high-power density and a time-domain hysteresis based highly integrated buck converter. A multi-phase time-domain comparator is proposed in this work for implementing the hysteresis control, thereby achieving a process scaling friendly highly digital design. A higher-order LC notch filter along with a flying capacitor which couples the input and output voltage ripple is implemented. The power stage operates at 500 MHz and can deliver a maximum power of 1.0 W and load current of 1.67 A, while occupying 1.21 mm2 active die area. Thus achieving a power density of 0.867 W/mm2 and current density of 1.377 A/mm2. The peak efficiency obtained is 71% at 780 mA of load current. The power stage with the additional off-chip LC is utilized to design a highly integrated current mode hysteretic buck converter operating at 180 MHz. It achieves 20 ns of settling and 2-5 ns of rise/fall time for reference tracking.
The second part of the dissertation discusses an integrated low voltage switched-capacitor based power sensor, to measure the output power of a DC-DC boost converter. This approach results in a lower complexity, area, power consumption, and a lower component count for the overall PV MPPT system. Designed in a 180 nm CMOS process, the circuit can operate with a supply voltage of 1.8 V. It achieves a power sense accuracy of 7.6%, occupies a die area of 0.0519 mm2, and consumes 0.748 mW of power.
ContributorsSingh, Shrikant (Author) / Kiaei, Sayfe (Thesis advisor) / Bakkaloglu, Bertan (Thesis advisor) / Kitchen, Jennifer (Committee member) / Song, Hongjiang (Committee member) / Arizona State University (Publisher)
Created2019
Description
The availability of bulk gallium nitride (GaN) substrates has generated great interest in the development of vertical GaN-on-GaN power devices. The vertical devices made of GaN have not been able to reach their true potential due to material growth related issues. Power devices typically have patterned p-n, and p-i junctions in lateral, and vertical direction relative to the substrate. Identifying the variations from the intended layer design is crucial for failure analysis of the devices. A most commonly used dopant profiling technique, secondary ion mass spectroscopy (SIMS), does not have the spatial resolution to identify the dopant distribution in patterned devices. The possibility of quantitative dopant profiling at a sub-micron scale for GaN in a scanning electron microscope (SEM) is discussed. The total electron yield in an SEM is shown to be a function of dopant concentration which can potentially be used for quantitative dopant profiling.
Etch-and-regrowth is a commonly employed strategy to generate the desired patterned p-n and p-i junctions. The devices involving etch-and-regrowth have poor performance characteristics like high leakage currents, and lower breakdown voltages. This is due to damage induced by the dry etching process, and the nature of the regrowth interface, which is important to understand in order to address the key issue of leakage currents in etched and regrown devices. Electron holography is used for electrostatic potential profiling across the regrowth interfaces to identify the charges introduced by the etching process. SIMS is used to identify the impurities introduced at the interfaces due to etch-and-regrowth process.
Etch-and-regrowth is a commonly employed strategy to generate the desired patterned p-n and p-i junctions. The devices involving etch-and-regrowth have poor performance characteristics like high leakage currents, and lower breakdown voltages. This is due to damage induced by the dry etching process, and the nature of the regrowth interface, which is important to understand in order to address the key issue of leakage currents in etched and regrown devices. Electron holography is used for electrostatic potential profiling across the regrowth interfaces to identify the charges introduced by the etching process. SIMS is used to identify the impurities introduced at the interfaces due to etch-and-regrowth process.
ContributorsAlugubelli, Shanthan Reddy (Author) / Ponce, Fernando A. (Thesis advisor) / McCartney, Martha (Committee member) / Newman, Nathan (Committee member) / Zhao, Yuji (Committee member) / Arizona State University (Publisher)
Created2019
Description
Silicon photovoltaics is the dominant contribution to the global solar energy production. As increasing conversion efficiency has become one of the most important factors to lower the cost of photovoltaic systems, the idea of making a multijunction solar cell based on a silicon bottom cell has attracted broad interest. Here the potential of using dilute nitride GaNPAs alloys for a lattice-matched 3-terminal 2-junction Si-based tandem solar cell through multiscale modeling is investigated. To calculate the electronic band structure of dilute nitride alloys with relatively low computational cost, the sp^3 d^5 s^* s_N tight-binding model is chosen, as it has been demonstrated to obtain quantitatively correct trends for the lowest conduction band near Γ, L, and X for dilute-N GaNAs. A genetic algorithm is used to optimize the sp^3 d^5 s^* tight-binding model for pure GaP and GaAs for their optical properties. Then the optimized sp^3 d^5 s^* s_N parametrizations are obtained for GaNP and GaNAs by fitting to experimental bandgap values. After that, a virtual crystal approach gives the Hamiltonian for GaNPAs alloys. From their tight-binding Hamiltonian, the first-order optical response functions of dilute nitride GaNAs, GaNP, and GaNPAs are calculated. As the N mole fraction varies, the calculated critical optical features vary with the correct trends, and agree well with experiment. The calculated optical properties are then used as input for the solar device simulations based on Silvaco ATLAS. For device simulation, a bottom cell model is first constructed to generate performance results that agree well with a demonstrated high-efficiency Si heterojunction interdigitated back contact (IBC) solar cell reported by Kaneka. The front a-Si/c-Si interface is then replaced by a GaP/Si interface for the investigation of the sensitivity of the GaP/Si interface to interface defects in terms of degradation of the IBC cell performance, where we find that an electric field that induces strong band bending can significantly mitigate the impact of the interfacial traps. Finally, a lattice-matched 3-terminal 2-junction tandem model is built for performance simulation by stacking a dilute nitride GaNP(As) cell on the Si IBC cell connected through a GaP/Si interface. The two subcells operate quasi-independently. In this 3-terminal tandem model, traps at the GaP/Si interface still significantly impact the performance of the Si subcell, but their effects on the GaNP subcell are relatively small. Assuming the interfacial traps are well passivated, the tandem efficiency surpasses that of a single-junction Si cell, with values close to 33% based on realistic parameters.
ContributorsZou, Yongjie (Author) / Goodnick, Stephen M. (Thesis advisor) / Honsberg, C. (Christiana B.) (Committee member) / King, Richard R. (Committee member) / Vasileska, Dragica (Committee member) / Arizona State University (Publisher)
Created2019
Description
State-of-the-art automotive radars use multi-chip Frequency Modulated Continuous Wave (FMCW) radars to sense the environment around the car. FMCW radars are prone to interference as they operate over a narrow baseband bandwidth and use similar radio frequency (RF) chirps among them. Phase Modulated Continuous Wave radars (PMCW) are robust and insensitive to interference as they transmit signals over a wider bandwidth using spread spectrum technique. As more and more cars are equipped with FMCW radars illuminate the same environment, interference would soon become a serious issue. PMCW radars can be an effective solution to interference in the noisy FMCW radar environment. PMCW radars can be implemented in silicon as System-on-a-chip (SoC), suitable for Multiple-Input-Multiple-Output (MIMO) implementation and is highly programmable. PMCW radars do not require highly linear high frequency chirping oscillators thus reducing the size of the final solution.
This thesis aims to present a behavior model for this promising Digitally modulated radar (DMR) transceiver in Simulink/Matlab. The goal of this work is to create a model for the electronic system level framework that simulates the entire system with non-idealities. This model includes a Top Down Design methodology to understand the requirements of the individual modules’ performance and thus derive the specifications for implementing the real chip. Back annotation of the actual electrical modules’ performance to the model closes the design process loop. Using Simulink’s toolboxes, a passband and equivalent baseband model of the system is built for the transceiver with non-idealities of the components built in along with signal processing routines in Matlab. This model provides a platform for system evaluation and simulation for various system scenarios and use-cases of sensing using the environment around a moving car.
This thesis aims to present a behavior model for this promising Digitally modulated radar (DMR) transceiver in Simulink/Matlab. The goal of this work is to create a model for the electronic system level framework that simulates the entire system with non-idealities. This model includes a Top Down Design methodology to understand the requirements of the individual modules’ performance and thus derive the specifications for implementing the real chip. Back annotation of the actual electrical modules’ performance to the model closes the design process loop. Using Simulink’s toolboxes, a passband and equivalent baseband model of the system is built for the transceiver with non-idealities of the components built in along with signal processing routines in Matlab. This model provides a platform for system evaluation and simulation for various system scenarios and use-cases of sensing using the environment around a moving car.
ContributorsKalyan, Prassana (Author) / Bakkaloglu, Bertan (Thesis advisor) / Kitchen, Jennifer (Thesis advisor) / Garrity, Douglas (Committee member) / Arizona State University (Publisher)
Created2019
Description
Learning analytics application is evolving into a student-facing solution. Student-facing learning analytics dashboards (SFLADs), as one popular application, occupies a pivotal position in online learning. However, the application of SFLADs faces challenges due to teacher-centered and researcher-centered approaches. The majority of SFLADs report student learning data to teachers, administrators, and researchers without direct student involvement in the design of SFLADs. The primary design criteria of SFLADs is developing interactive and user-friendly interfaces or sophisticated algorithms that analyze the collected data about students’ learning activities in various online environments. However, if students are not using these tools, then analytics about students are not useful. In response to this challenge, this study focuses on investigating student perceptions regarding the design of SFLADs aimed at providing ownership over learning. The study adopts an approach to design-based research (DBR; Barab, 2014) called the Integrative Learning Design Framework (ILDF; Bannan-Ritland, 2003). The theoretical conjectures and the definition of student ownership are both framed by Self-determination theory (SDT), including four concepts of academic motivation. There are two parts of the design in this study, including prototypes design and intervention design. They are guided by a general theory-based inference which is student ownership will improve student perceptions of learning in an autonomy-supportive SFLAD context. A semi-structured interview is used to gather student perceptions regarding the design of SFLADs aimed at providing ownership over learning.
ContributorsLi, Siyuan (Author) / Zuiker, Steven (Thesis advisor) / Cunningham, James (Committee member) / Lande, Micah (Committee member) / Arizona State University (Publisher)
Created2019