Theses and Dissertations at Montana State University (MSU)

Permanent URI for this collectionhttps://scholarworks.montana.edu/handle/1/733

Browse

Search Results

Now showing 1 - 10 of 26

Sparse descriptors for whole graph embedding and dictionary based feature ranking
(Montana State University - Bozeman, College of Engineering, 2024) Liyanage, Liyanpathige Kaveen Gayasara; Chairperson, Graduate Committee: Brad Whitaker
Graph representation has gained wide popularity as a data representation method in many applications. Unfortunately, most data processing techniques cannot be applied directly to a graph structure. Therefore, graph embedding methods are frequently used to convert graphs to vectors. While such methods are essential in standard data processing pipelines, they often result in complicated, nonlinear, and high-dimensional mappings. The goal of this dissertation is to utilize sparse dictionary learning techniques in the context of graph embedding. In contrast to traditional graph embedding methods, sparse representations are linear by design. This linearity also leads to intuition, since the building blocks of a sparse dictionary are directly related to the input space. Despite the potential advantages of sparse processing and the ubiquitousness of sparsity in other signal processing domains, its applications in graph embedding are not well studied. This dissertation consists of three main tasks. First, a novel sparse graph descriptor algorithm is presented, inspired by the Graph2Vec graph embedding algorithm. Second, sparse representation-based feature ranking metrics are deployed to identify important sub- tree structures of the graphs that can be used to define a dictionary. The developed embedding algorithm and feature-ranking metrics are compared to existing graph embedding methods and feature-ranking algorithms on several typical benchmark graph datasets. Finally, these sparse representation-based techniques are applied to control flow graphs of binary files to detect malware, showing the utility of the developed algorithms.
A micropulse differential absorption LIDAR for temperature profiling
(Montana State University - Bozeman, College of Engineering, 2024) Cruikshank, Owen Daniel; Chairperson, Graduate Committee: Kevin S. Repasky
This dissertation describes the development and testing of a Differential Absorption Lidar (DIAL) technique to retrieve temperature. This work showcases the accuracy of temperature retrievals using a perturbative technique, combining a DIAL measurement of a temperature-sensitive oxygen (O 2 ) absorption profile with a high spectral resolution lidar (HSRL) measurement of the backscatter ratio profile near 770 nm. This dissertation also introduces advancements to the DIAL temperature instrument and retrieval. First, the spectroscopic model used to represent absorption of light by O 2 has been enhanced, via a more complete physical representation, improving measurement accuracy. Second, the error estimation and masking have been developed using the bootstrapping technique. Third, a comparison of temperature profiles from the instrument with collocated radiosondes, evaluating the accuracy of updated measurements is performed. Finally, modeling of the lidar overlap function and atmospheric propagation and temperature retrieval allows for the modeling of sources of bias. A test of the DIAL temperature instrument was performed by NSF NCAR (National Center for Atmospheric Research) in Fort Collins CO in 2021. The data were collected between June 17 and August 20, 2021. Radiosonde comparisons were available for comparison. Results show temperature comparisons with radiosonde that have a bias of about 1 oC and a standard deviation of about 3 oC. Another test of the DIAL temperature instrument was performed at Montana State University. The laboratory-based lidar instrument was operated over a six-month period between April 21, 2022 and September 22, 2022. During this time, we launched 40 radiosondes, providing reference data to validate the accuracy of the DIAL-based temperature profiles. The results indicate that DIAL-based temperature retrievals are within + or - 2.5 oC between 0.4 km and 3 km (3.5 km) range during daytime (nighttime) operation, using a 300 m range resolution and a 60 min averaging time.
Novel approach to fault tolerance in space computers leveraging the RISC-V architecture
(Montana State University - Bozeman, College of Engineering, 2023) Major, Chris Michel; Chairperson, Graduate Committee: Brad Whitaker
As the aerospace industry continues to accelerate in growth and mission frequency, the demand for high-performance computers that can withstand radiation environments has become a critical need within the field. Traditional space computing systems rely on specialized and complex means of radiation resistance, but more modern systems seek to implement redundant components to mitigate radiation effects. This dissertation presents a novel approach to radiation-tolerant space computing, based on Montana State University's RadPC program, by developing a resilient architecture leveraging the open-source RISC- V processor. The architecture discussed is designed to withstand radiation environments and engage repairs for damaged sections using commercial, off-the-shelf components - without requiring radiation-hardened fabrication processes or specialized manufacturing. This dissertation discusses the design and performance of the components required to ensure radiation resilience in the system, reconfigure compromised processors in the event of damage, and provides a characterization of the system's overall performance in various space environments.
Using instruction code obfuscation to defeat malware attacks
(Montana State University - Bozeman, College of Engineering, 2023) Running Crane, Tristan Tanner; Chairperson, Graduate Committee: Brock LaMeres
This thesis investigates the use of obfuscated instruction codes within a redundant, RISC-V computer architecture to detect and defeat malware injected cyberattacks. The system abstracts the obfuscation from the user so that the system looks like a single processor edge computer. The system was tested using real-time sensor data coming from a camera while image processing algorithms were performed. The results of this thesis contribute to the body of knowledge on how to keep edge computers used in critical applications operational during a cyber attack.
Hardware and software development for implementation of fast and safe charging of commercial lihtium-ion batteries
(Montana State University - Bozeman, College of Engineering, 2023) Hedding, Noah Robert; Chairperson, Graduate Committee: Hongwei Gao
From single cells in handheld electronics to enormous packs in battery electric vehicles (BEV), batteries govern modern life. Lithium ion batteries (LIB) present the best available commercially available products for these applications; they have the highest energy densities and can output currents many times their capacity. But safely charging LIBs requires a slow and detailed process which is typically unacceptable for use in BEV and other rugged handheld devices; therefore, decreasing the required charging time would be greatly beneficial. Fast charging methods do present dangers and concerns. Unmonitored fast charging of LIBs allows for the potential of lithium plating where the lithium ions within the cell are converted to metallic lithium at the battery anode. Lithium plating can remove these ions from the charging and discharging process causing reductions in battery capacity. The metallic lithium structures formed also present the dangers of short circuit and thermal runaway. In this thesis, a charging protocol is developed using equivalent circuit models and experimentation with the goal of the elimination of lithium plating. First, equivalent models of a test cell were determined and validated. Then, this test cell was used to find the fast charging protocol both experimentally and through the use the equivalent circuit elements. Custom power electronics and software were then developed to implement the proposed charging protocol on commercial LIBs for 350 cycles. The results of this experiment show that the charging protocol did not create noticeable lithium plating while decreasing the charging time required by a typical constant current - constant voltage (CC/CV) from 50 minutes to 29 minutes. The proposed charging protocol decreased the charging time without stressing the LIB beyond its set limitation.
Injection attack immunity using redundant heterogeneous processing cores
(Montana State University - Bozeman, College of Engineering, 2023) Barney, Colter Ross; Chairperson, Graduate Committee: Brock LaMeres
Technology is an integral part of modern society. Devices such as smart lights, locks and appliances are becoming more commonplace. This class of devices are called embedded systems. Embedded systems can be targeted by malicious cyber attacks just as a normal computer can. Unfortunately, many techniques used to secure and protect normal computers do not work on embedded systems. New security techniques must be developed and designed to protect embedded systems. This paper investigates using physically diverse processing cores to defeat cyber attacks in real time. Diverse processing cores were implemented using reconfigurable hardware devices called FPGAs. The use of FPGAs allows diverse cores to be utilized, without losing the benefits gained from standardized processors. The cores implemented were based on a commercial processor made by Texas Instruments (TI). Modeling the diverse cores after a commercial processor enables the cores to utilize development tools created for TI's processor. A complete system was built using diverse processors to prove the feasibility and usability of secure embedded systems. The cores were used to control a realistic embedded system application. While operating, the cores were subjected to a cyber attack, and they were able to nullify the attack. An identical setup was created using the commercially available processor. Attacking the commercial processor compromised the application and reinforced the need for secure systems. The techniques investigated and utilized in this paper can be expanded to increase security in the many embedded systems that have become an essential part of modern lifestyles.
Fault injection system for FPGA-based space computers
(Montana State University - Bozeman, College of Engineering, 2023) Austin, Hezekiah Ajax; Chairperson, Graduate Committee: Brock LaMeres
Abstract: Simulation of radiation effects in aerospace computers is a key testing and verification component to space operations. Contemporary computer architectures utilizing Field Programmable Gate Arrays (FPGA) requires particular focus in testing the configuration memory of the device for faults that cannot be recovered using traditional strategies. Faults in the configuration memory propagate to the hardware settings of the FPGA, changing the implemented logic circuit functionality. The effects of faults in the configuration memory are unpredictable, limiting the effectiveness of computer simulation and analysis. Therefore, designers of FPGA-based aerospace computers prefer to physically induce faults in the configuration memory to measure their impact. This allows the results of configuration memory fault injection used to classify faults occurring during space operation. The process is difficult to implement as the FPGA configuration memory is large, often undocumented, and the injection process is tedious when done manually. This paper presents the results of the deployment of two FPGA-based aerospace computers payloads to the International Space Station and the subsequently developed process for configuration memory fault injection. The injections are designed to simulate errors caused by radiation strikes to the computer hardware. These injections were performed on duplicate hardware to the RadPC payloads that operated on the ISS and was bombardment by real radiation. This provided the ability to see if the ground-based injection was correlated to real flight data. The developed process is able to inject single bit faults, which represents the majority of faults observed in configuration memory for space applications, and continuous injection, which stress tests the aerospace computer's recovery capability. Depending on the effects of the injected fault, the error is marked as either repairable, nonrepairable and propagating, or nonrepairable and nonpropagating. The result of this testing illustrates the key components in the implemented computer architecture which are vulnerable to faults in the configuration memory. Vulnerable components include the softcores, voter components, and the input logic. The process allows these key components to be isolated for further testing and the comparison of payload results to configuration memory testing on the ground.
Spectral processing for algae monitoring and mapping (SPAMM): remote sensing methodologies for river ecology
(Montana State University - Bozeman, College of Engineering, 2024) Logan, Riley Donovan; Chairperson, Graduate Committee: Joseph A. Shaw; This is a manuscript style paper that includes co-authored chapters.
Inland water quality is a growing concern to public health, riparian ecosystems, and recreational uses of our waterways. Many modern water quality programs include measures of the presence and abundance of harmful and nuisance algae. In southwestern Montana, large blooms of the nuisance algae, Cladophora glomerata, have become common in the Upper Clark Fork River due to a combination of warming water temperatures, naturally high phosphorus levels, and an influx of contaminants through wastewater and anthropogenic activity along its banks. To improve understanding of bloom dynamics, such as algal biomass and percent algae cover, and their effects on water quality, a UAV-based hyperspectral imaging system was used to monitor several locations along the Upper Clark Fork River. Image data were collected across the spectral range of 400 - 1000 nm with 2.1 nm spectral resolution during field sampling campaigns across the entirety of the project, beginning in 2019 and ending in 2023. In this dissertation, methodologies for monitoring water quality were developed. These methods include estimating benthic algal pigment abundance using spectral band ratios achieving R 2 values of up to 0.62 for chlorophyll alpha and 0.96 for phycocyanin; creating spatial algae distribution maps and estimating percent algae cover using machine learning classification algorithms with accuracies greater than 99%; combining spatial algae distribution maps and improved pigment estimation using machine learning regression algorithms for creating chlorophyll alpha abundance maps, achieving an R 2 of 0.873, while also comparing abundance values to Montana water quality thresholds; and identifying salient wavelengths for monitoring and mapping algae to inform the design of a low-cost and compact multispectral imager. Throughout all field campaigns, significant spatial variations in algal growth within each river reach and frequent violations of current water quality standards were observed, demonstrating the need for high-spatial resolution monitoring techniques to be incorporated in current water quality monitoring programs.
MEMS cantilevers for dynamic strain studies of 2D materials
(Montana State University - Bozeman, College of Engineering, 2024) Heris, Masoud Hakimi; Chairperson, Graduate Committee: David L. Dickensheets
Photoluminescence properties of transition metal dichalcogenide (TMDC) 2D materials are known to vary sensitively with strain. There have been a few attempts to create tunable mechanical strain using silicon MEMS techniques, but these initial devices were limited in the maximum strain achievable, ability to operate in cryogenic environments, or ability to generate strain dynamically. This project is developing microelectromechanical systems (MEMS) for dynamic electronic control of external strain in 2D materials. These electrostatically actuated cantilevers are intended for either vertical or lateral continuous deflection or full snap-down (or to the side for lateral case), with up to 2 micron of vertical travel and a range of 2-10 micron for the lateral cantilever. According to geometrical calculations, these cantilevers could deliver up to 5.9% strain with vertical motion, and up to 33.3% strain with lateral motion. Such strain is more than sufficient to transition some 2D semiconductors from a direct to an indirect bandgap material and fully modulate light emission processes. In this project test coupons with vertically and laterally actuated cantilevers were created using silicon-on-insulator wafers. The benefit of making use of SOI wafers is a thick device layer leading to robust cantilevers that will survive the flake transfer, and an easy fabrication process compared with previous technologies of MEMS fabrication. It has been shown that thicker device layer SOI wafers could reduce the stress-induced elevation of the cantilevers above the substrate plane after release etch. BOX layer thickness of 2 micron allows for reasonably low actuation voltage while keeping the vertical travel long enough for inducing considerable strain. In order to eventually operate the chip in the cryostat for cryogenic characterization of the strained material, highly Boron-doped SOI wafers have been purchased. Gold assisted exfoliation and dry transfer techniques have been successfully used for transferring 2D TMDC over the MEMS platform for PL measurements. Based on PL measurements and comparing with literature data reported about WSe2, we could confirm about 0.031% of strain that could be induced at room temperature with MEMS actuation. This is lower than the geometric lengthening we predict, which we attribute to poor adhesion and sliding of the flake over the cantilever. This suggests that future work may profit from a mechanism to clamp the flake on the cantilever as well as the anchored surrounding substrate, in order to increase the amount of strain the MEMS can deliver to the flake.
Data-driven approaches for distribution grid modernization: exploring state estimaion, pseudo-measurement generation and false data detection
(Montana State University - Bozeman, College of Engineering, 2023) Radhoush, Sepideh; Chairperson, Graduate Committee: Brad Whitaker
Distribution networks must be regularly updated to enhance their performance and meet customer electricity requirements. Advanced technologies and infrastructure--including two- way communication, smart measuring devices, distributed generations in various forms, electric vehicles, variable loads, etc.--have been added to improve the overall efficiency of distribution networks. Corresponding to these new features and structures, the continuous control and monitoring of distribution networks should be intensified to keep track of any modifications to the distribution network performance. Distribution system state estimation has been introduced for real-time monitoring of distribution networks. State estimation calculations are highly dependent on measurement data which are collected from measurement devices in distribution networks. However, the installation of measurement devices is not possible at all buses to ensure the distribution network is fully observable. To address the lack of real measurements, pseudo- measurements are produced from historical load and generation data. Available measurements, along with physical distribution network topology, are fed into a state estimation algorithm to determine system state variables. Then, state estimation results are sent to a control center for further processing to enhance distribution network operation. However, the accuracy of state estimation results could be degraded by false data injection attacks on measurement data. If these attacks are not detected, distribution network operation could be significantly influenced. Different methods have been developed to enhance a distribution network operation and management. Machine learning approaches have also been identified to be beneficial in solving different types of problems in a power grid. In this dissertation, machine learning is applied to three areas of distribution systems: generating pseudo-measurements, performing distribution system state estimation calculations, and detecting false data injection attacks on measurement data. In addition to addressing these areas individually, machine learning is used to simultaneously perform distribution system state estimation calculation and false data injection attack detection. This is done by taking advantage of conventional and smart measurement data at different time scales. The results reveal that the operation and performance of a distribution network are improved using machine learning algorithms, leading to more effective power grid modernization.

Theses and Dissertations at Montana State University (MSU)

Browse

Filters

Settings

Sort By

Results per page

Search Results