College of Engineering (COE)

Permanent URI for this community

https://hdl.handle.net/10919/5539

Note: The Department of Biological Systems Engineering is listed within the College of Agriculture and Life Sciences (CALS).

Browse

Now showing 1 - 20 of 270

2 kV, 0.7 mΩ·cm² Vertical Ga₂O₃ Superjunction Schottky Rectifier with Dynamic Robustness
Qin, Yuan; Porter, Matthew; Xiao, Ming; Du, Zhonghao; Zhang, Hongming; Ma, Yunwei; Spencer, Joseph; Wang, Boyan; Song, Qihao; Sasaki, Kohei; Lin, Chia-Hung; Kravchenko, Ivan; Briggs, Dayrl P.; Hensley, Dale K.; Tadjer, Marko; Wang, Han; Zhang, Yuhao (IEEE, 2023)
We report the first experimental demonstration of a vertical superjunction device in ultra-wide bandgap (UWBG) Ga2O3. The device features 1.8 μm wide, 2×1017 cm-3 doped n-Ga2O3 pillars wrapped by the charge-balanced p-type nickel oxide (NiO). The sidewall NiO is sputtered through a novel self-align process. Benefitted from the high doping in Ga2O3, the superjunction Schottky barrier diode (SJ-SBD) achieves a ultra-low specific on-resistance (RON,SP) of 0.7 mΩ·cm2 with a low turn-on voltage of 1 V and high breakdown voltage (BV) of 2000 V. The RON,SP~BV trade-off is among the best in all WBG and UWBG power SBDs. The device also shows good thermal stability with BV > 1.8 kV at 175 oC. In the unclamped inductive switching tests, the device shows a dynamic BV of 2.2 kV and no degradation under 1.7 kV repetitive switching, verifying the fast acceptor depletion in NiO under dynamic switching. Such high-temperature and switching robustness are reported for the first time in a heterogeneous superjunction. These results show the great potential of UWBG superjunction power devices.
2023 Engineering Mechanics Research Symposium
(Virginia Tech, 2023-03)
The annual Engineering Mechanics Spring Symposium featured lecture and poster presentations from both students and faculty members.
The 4th Workshop on Localization vs. Internationalization: Creating an International Survey on Automotive User Interfaces
Stojmenova, Kristina; Lee, Seul Chan; De Oliveira Faria, Nayara; Schroeter, Ronald; Jeon, Myounghoon (ACM, 2022-09-17)
International surveys tend to collect data on attitudes, values and behaviors towards a specific topic from users from multiple countries, providing an insight on the differences and similarities across nations, cultures or geo-political structures. Consequently, international surveys provide important information about the diversity of the user's needs, values and preferences, which have to be taken into consideration when creating products and services as widely used as the personal automobile. The workshop will focus on the design and development of an international survey on automotive user interfaces on a global scale. It will try to identify the most important aspects related to automotive user interfaces, which should be addressed in the survey. It will also prepare a strategy for its international distribution and create a plan for comprehensive data collection. Lastly, it will try to outline venues and communication channels for the survey dissemination, with the goal of achieving wide visibility.
5G Opportunities in Warehousing
George, Roshan; Cherbaka, Natalie; Ellis, Kimberly (2023)
As capabilities of fifth generation wireless technology (5G) improve, adoption will go beyond current urban cellular networks into industrial settings enabling the IoT landscape. 5G primarily delivers value by enhancing mobile broadband through ultra-reliable, low-latency signals and massive machine-type communications. With the concurrent development of 5G and industrial automation, replacing Wi-Fi and LTE services with 5G networks offers an opportunity to enhance scheduling, latency, jitters, and redundancy in demanding applications. Additionally, the equipment redesigns and upgrades to operate in 5G will pave the way for innovation in operational strategies previously constrained by network capabilities. In this paper, we consider the warehouse operations and functions that are most likely to benefit from 5G adoption. The areas 5G will impact in warehousing are robotic operations, such as AGVs/AMRs; augmented reality devices for picking, training, and maintenance; inventory management through real time asset tracking; equipment battery life from network slicing; and data security. In general, the capacity and low-latency available through 5G will support continuous data transfer that is sufficient to support real-time analytics and decision-making. Knowing which functions will benefit most from 5G will provide strategic guidance for upgrading equipment and operations and aid in developing the factory of the future.
Accelerated Corrosion Testing of ASTM A1010 Stainless Steel
Hebdon, Matthew H.; Groshek, Isaac (American Institute of Steel Construction, 2018-04-11)
ASTM A1010 (recently adopted as ASTM A709 Gr50CR) is a material which has advantageous corrosion properties. It is a low-grade stainless steel which forms a protective patina and has been marketed as an alternative to other bridge steels and corrosion protection methods due to its corrosion resistance in highly corrosive environments. However, the material is currently available in plate form only, and several of the applications in the United States were required to use alternative materials when constructing and connecting secondary members to the A1010 plate girders. This paper addresses the corrosion behavior of A1010 in several different details relating to recent applications in the US. An accelerated corrosion study was performed which simulated a highly corrosive environment typical of the environment justifying the use of A1010. The research investigated the resulting galvanic corrosion and its effect on the corrosion rate of A1010 plates, several different common bridge steels, and typical fastener materials. In addition, common surface preparation methods were evaluated for their aesthetic effect during patina formation.
ADOC: Automatically Harmonizing Dataflow Between Components in Log-Structured Key-Value Stores for Improved Performance
Yu, Jinghuan; Noh, Sam H.; Choi, Young-ri R.; Xue, Chun Jason (Usenix Association, 2023)
Log-Structure Merge-tree (LSM) based Key-Value (KV) systems are widely deployed. A widely acknowledged problem with LSM-KVs is write stalls, which refers to sudden performance drops under heavy write pressure. Prior studies have attributed write stalls to a particular cause such as a resource shortage or a scheduling issue. In this paper, we conduct a systematic study on the causes of write stalls by evaluating RocksDB with a variety of storage devices and show that the conclusions that focus on the individual aspects, though valid, are not generally applicable. Through a thorough review and further experiments with RocksDB, we show that data overflow, which refers to the rapid expansion of one or more components in an LSM-KV system due to a surge in data flow into one of the components, is able to explain the formation of write stalls. We contend that by balancing and harmonizing data flow among components, we will be able to reduce data overflow and thus, write stalls. As evidence, we propose a tuning framework called ADOC (Automatic Data Overflow Control) that automatically adjusts the system configurations, specifically, the number of threads and the batch size, to minimize data overflow in RocksDB. Our extensive experimental evaluations with RocksDB show that ADOC reduces the duration of write stalls by as much as 87.9% and improves performance by as much as 322.8% compared with the auto-tuned RocksDB. Compared to the manually optimized state-of-the-art SILK, ADOC achieves up to 66% higher throughput for the synthetic write-intensive workload that we used, while achieving comparable performance for the real-world YCSB workloads. However, SILK has to use over 20% more DRAM on average.
Advanced Boundary Simulations of an Aeroacoustic and Aerodynamic Wind Tunnel
Szőke, Máté; Devenport, William J.; Borgoltz, Aurelien; Roy, Christopher J.; Lowe, K. Todd (2021-05-25)
This study presents the first 3D two-way coupled fluid structure interaction (FSI) simulation of a hybrid anechoic wind tunnel (HAWT) test section with modeling all important effects, such as turbulence, Kevlar wall porosity and deflection, and reveals for the first time the complete 3D flow structure associated with a lifting model placed into a HAWT. The Kevlar deflections are captured using finite element analysis (FEA) with shell elements operated under a membrane condition. Three-dimensional RANS CFD simulations are used to resolve the flow field. Aerodynamic experimental results are available and are compared against the FSI results. Quantitatively, the pressure coefficients on the airfoil are in good agreement with experimental results. The lift coefficient was slightly underpredicted while the drag was overpredicted by the CFD simulations. The flow structure downstream of the airfoil showed good agreement with the experiments, particularly over the wind tunnel walls where the Kevlar windows interact with the flow field. A discrepancy between previous experimental observations and juncture flow-induced vortices at the ends of the airfoil is found to stem from the limited ability of turbulence models. The qualitative behavior of the flow, including airfoil pressures and cross-sectional flow structure is well captured in the CFD. From the structural side, the behavior of the Kevlar windows and the flow developing over them is closely related to the aerodynamic pressure field induced by the airfoil. The Kevlar displacement and the transpiration velocity across the material is dominated by flow blockage effects, generated aerodynamic lift, and the wake of the airfoil. The airfoil wake increases the Kevlar window displacement, which was previously not resolved by two-dimensional panel-method simulations. The static pressure distribution over the Kevlar windows is symmetrical about the tunnel mid-height, confirming a dominantly two-dimensional flow field.
Adversarial Unlearning of Backdoors via Implicit Hypergradient
Zeng, Yi; Chen, Si; Park, Won; Mao, Morley; Jin, Ming; Jia, Ruoxi (2022)
We propose a minimax formulation for removing backdoors from a given poisoned model based on a small set of clean data. This formulation encompasses much of prior work on backdoor removal. We propose the Implicit Bacdoor Adversarial Unlearning (I-BAU) algorithm to solve the minimax. Unlike previous work, which breaks down the minimax into separate inner and outer problems, our algorithm utilizes the implicit hypergradient to account for the interdependence between inner and outer optimization. We theoretically analyze its convergence and the generalizability of the robustness gained by solving minimax on clean data to unseen test data. In our evaluation, we compare I-BAU with six stateof- art backdoor defenses on seven backdoor attacks over two datasets and various attack settings, including the common setting where the attacker targets one class as well as important but underexplored settings where multiple classes are targeted. I-BAU’s performance is comparable to and most often significantly better than the best baseline. Particularly, its performance is more robust to the variation on triggers, attack settings, poison ratio, and clean data size. Moreover, I-BAU requires less computation to take effect; particularly, it is more than 13X faster than the most efficient baseline in the single-target attack setting. Furthermore, it can remain effective in the extreme case where the defender can only access 100 clean samples—a setting where all the baselines fail to produce acceptable results.
AI-Powered Real-Time Channel Awareness and 5G NR Radio Access Network Scheduling Optimization
Wang, Ying; Gorski, Adam; DaSilva, Luiz A. (IEEE, 2021-04-19)
As with any other wireless technology, 5G is not immune to jamming. To achieve consistent performance, network resource scheduling must be optimized in a way that reacts to jamming in the NR channel environment. This paper presents a cognitive system for real-time Channel Awareness and Radio Access Network (RAN) Scheduling (CARS) optimization based on multi-dimensional temporal machine learning models. Our system automatically detects and classifies jamming in the channel environment and optimizes scheduling based on classification results and collected link parameters. Based on over-the-air (OTA) experiments, detection and classification time is less than 0.8 seconds, which enables real-time optimization. The system is evaluated and verified for OTA experimentation through integration to our end-to-end NR system. An Automated Jamming Module (AJM) is designed and implemented. Connecting the AJM to our NR system enables a comprehensive evaluation environment for our Jamming Detection and Classification Model (JDCM) and Modulation and Coding Scheme optimization model. The improvement in connection resiliency against Control Resource Set jamming is proof of the CARS concept for real-time channel awareness and scheduling optimization. Depending on channel conditions, CARS achieves a 30% or higher improvement in NR system throughput.
Airport Scheduling and Operational Performance: A Clustering Analysis of Airport Response to COVID-19
Alsalous, Osama; Hotle, Susan (American Institute of Aeronautics and Astronautics, 2023-06)
In early 2020, the Coronavirus disease 2019 (COVID-19) pandemic started and forced air travel demand to decrease sharply in most parts of the world due to travel restrictions that were put in place to limit the spread of the virus. The pandemic also impacted capacity due to reasons such as workforce social distancing, days when Air Traffic Control (ATC) facilities were shut down due to COVID cases, and financial challenges due to the decreased demand. The reduced demand created a unique challenge in the system since capacity exceeded demand by very large margins in the NAS, however, delays in the system did not fall to zero despite the sharp drop in demand. This study analyzed operations at 77 United States (US) airports to compare and contrast their responses to the COVID-19 pandemic in terms of capacity, throughput, and the resulting operational performance. We evaluate the response of airports to the initial shock event during 2020 in addition to the recovery period that followed in 2021. The data showed a 67% decline in total operations at the lowest point during the pandemic. The impact during the shock time period varied greatly across the airports, ranging from a reduction of 14.8% at MEM to 81.5% at LGA. We performed a clustering analysis to study airports’ response to the COVID-19 pandemic. There was a number of airport characteristics that were correlated to the changes in airport metrics. For example, the data showed that being located in a multi-airport city was significantly correlated to the decrease in operations during the shock, however, it was not significant in the recovery trends. Our analysis showed that delays in the system did not change proportionately to the change in operations. Similarly, there were only minor improvements in punctuality, on-time flights at the ASPM 77 airports increased by 9.5% while operations declined by 52% during the shock event time period compared to pre-COVID. Part of this phenomenon was a result of schedule peaking which caused delays due to creating busy hours at the airports. This analysis can inform airport management when responding to future disruptive events, it provides insight into airport operational resiliency, response to disruption, and demand recovery patterns based on airport characteristics.
Application of surrogate models for performance-based evaluation of multi-story concrete buildings at early design
Zaker Esteghamati, Mohsen; Flint, Madeleine M.; Rodriguez-Marek, Adrian (2022)
Data incompleteness and uncertainty impede the application of performance-based design of structures at early design, which relies on data- and time-intensive numerical simulations. Early design is the most influential stage in a buildings' life cycle performance, hence neglecting quantitative methods to evaluate the design in preliminary stages can lead to missing on opportunities to improve building resiliency. This study presents a framework to implement surrogate models for supporting performance-based early design of concrete multi-story buildings. Five different surrogate models including multiple linear regression, random forest, extreme gradient boosting, support vector regression machines, and k-nearest neighbors are developed and compared to represent the seismic-induced structural loss of 720 generic concrete office buildings using early design parameters. Additionally, variance-based sensitivity is used to determine influential parameters for the best-performing model. The results show that extreme gradient boosting and support vector regression machines can be used to relate crude topology and design parameters to building seismic performance with reasonable accuracy.
Architecting a Cloud-native Data Analysis Application for ETDs
Chen, Yinlin; Fox, Edward A. (2018)
In this paper, we present a Cloud-native data analysis application and its architecture. This application was developed for librarians to explore useful information from the ETDs preserved in the Virginia Tech digital repository - VTechWorks. We realized the Cloud-native concepts by architecting a serverless architecture with microservices and managed services as backend, and deployed the entire application on Amazon Web Services (AWS). We detail our architecture strategies, decisions we made, and the best practices we followed. Furthermore, we share the lessons learned and cloud benefits we have gained. We believe that our proposed approach could be adopted by other ETD systems, e.g., NDLTD, and could be of benefit to the broader community.
An Architecture for Electronic Textiles
Jones, Mark T.; Martin, Thomas L.; Sawyer, Braden (ICST, 2008)
This paper makes a case for a communication architecture for electronic textiles (e-textiles). The properties and re- quirements of e-textile garments are described and analyzed. Based on these properties, the authors make a case for em- ploying wired, digital communication as the primary on- garment communication network. The implications of this design choice for the hardware architecture for e-textiles are discussed.
Augmented Reality’s Potential for Identifying and Mitigating Home Privacy Leaks
Cruz, Stefany; Danek, Logan; Liu, Shinan; Kraemer, Christopher; Wang, Zixin; Feamster, Nick; Huang, Danny Yuxing; Yao, Yaxing; Hester, Josiah (Internet Society, 2023)
Users face various privacy risks in smart homes, yet there are limited ways for them to learn about the details of such risks, such as the data practices of smart home devices and their data flow. In this paper, we present Privacy Plumber, a system that enables a user to inspect and explore the privacy “leaks” in their home using an augmented reality tool. Privacy Plumber allows the user to learn and understand the volume of data leaving the home and how that data may affect a user’s privacy— in the same physical context as the devices in question, because we visualize the privacy leaks with augmented reality. Privacy Plumber uses ARP spoofing to gather aggregate network traffic information and presents it through an overlay on top of the device in an smartphone app. The increased transparency aims to help the user make privacy decisions and mend potential privacy leaks, such as instruct Privacy Plumber on what devices to block, on what schedule (i.e., turn off Alexa when sleeping), etc. Our initial user study with six participants demonstrates participants’ increased awareness of privacy leaks in smart devices, which further contributes to their privacy decisions (e.g., which devices to block).
Augmenting Knowledge Transfer across Graphs
Mao, Yuzhen; Sun, Jianhui; Zhou, Dawei (IEEE, 2022-11)
Given a resource-rich source graph and a resource-scarce target graph, how can we effectively transfer knowledge across graphs and ensure a good generalization performance? In many high-impact domains (e.g., brain networks and molecular graphs), collecting and annotating data is prohibitively expensive and time-consuming, which makes domain adaptation an attractive option to alleviate the label scarcity issue. In light of this, the state-of-the-art methods focus on deriving domain-invariant graph representation that minimizes the domain discrepancy. However, it has recently been shown that a small domain discrepancy loss may not always guarantee a good generalization performance, especially in the presence of disparate graph structures and label distribution shifts. In this paper, we present TRANSNET, a generic learning framework for augmenting knowledge transfer across graphs. In particular, we introduce a novel notion named trinity signal that can naturally formulate various graph signals at different granularity (e.g., node attributes, edges, and subgraphs). With that, we further propose a domain unification module together with a trinity-signal mixup scheme to jointly minimize the domain discrepancy and augment the knowledge transfer across graphs. Finally, comprehensive empirical results show that TRANSNET outperforms all existing approaches on seven benchmark datasets by a significant margin.
Autofocusing in optical scanning holography
Kim, Taegeun; Poon, Ting-Chung (Optical Society of America, 2009-12-01)
We present autofocusing in optical scanning holography (OSH) with experimental results. We first record the complex hologram of an object using OSH and then create the Fresnel zone plate (FZP) that codes the object constant within the depth range of the object using Gaussian low-pass filtering. We subsequently synthesize a real-only spectrum hologram in which its phase term contains information about a distance parameter. Finally, we extract the distance parameter from the real-only spectrum hologram using fringe-adjusted filtering and the Wigner distribution. Using the extracted distance parameter, we reconstruct a three-dimensional image of the object from the complex hologram using digital convolution, which bypasses the conventional blind convolution to reconstruct a hologram. To the best of our knowledge, this is the first report with experimental results that autofocusing in OSH is possible without any searching algorithm or tracking process. (C) 2009 Optical Society of America
Automatic Metadata Extraction Incorporating Visual Features from Scanned Electronic Theses and Dissertations
Choudhury, Muntabir; Jayanetti, Himarsha R.; Wu, Jian; Ingram, William A.; Fox, Edward (IEEE, 2021-09-27)
Electronic Theses and Dissertations (ETDs) contain domain knowledge that can be used for many digital library tasks, such as analyzing citation networks and predicting research trends. Automatic metadata extraction is important to build scalable digital library search engines. Most existing methods are designed for born-digital documents such as GROBID, CERMINE, and ParsCit, so they often fail to extract metadata from scanned documents such as for ETDs. Traditional sequence tagging methods mainly rely on text-based features. In this paper, we propose a conditional random field (CRF) model that combines text-based and visual features. To verify the robustness of our model, we extended an existing corpus and created a new ground truth corpus consisting of 500 ETD cover pages with human validated metadata. Our experiments show that CRF with visual features outperformed both a heuristic baseline and a CRF model with only text-based features. The proposed model achieved 81.3%-96% F1 measure on seven metadata fields. The data and source code are publicly available on Google Drive1 and a GitHub repository2.
Automating context dependent gaze metrics for evaluation of laparoscopic surgery manual skills
Deng, Shiyu; Kulkarni, Chaitanya; Parker, Sarah J.; Barnes, Laura E.; Wang, Tianzi; Hartman-Kenzler, Jacob; Safford, Shawn; Lau, Nathan (2022-03)
AUTOPAGER: Auto-tuning Memory-Mapped I/O Parameters in Userspace
Youssef, Karim; Shah, Niteya; Gokhale, Maya; Pearce, Roger; Feng, Wu-chun (IEEE, 2022)
The exponential growth in dataset sizes has shifted the bottleneck of high-performance data analytics from the compute subsystem to the memory and storage subsystems. This bottleneck has led to the proliferation of non-volatile memory (NVM). To bridge the performance gap between the Linux I/O subsystem and NVM, userspace memory-mapped I/O enables application-specific I/O optimizations. Specifically, UMap, an open-source userspace memory-mapping tool, exposes tunable paging parameters to application users, such as page size and degree of paging concurrency. Tuning these parameters is computationally intractable due to the vast search space and the cost of evaluating each parameter combination. To address this challenge, we present Autopager, a tool for auto-tuning userspace paging parameters. Our evaluation, using five data-intensive applications with UMap, shows that Autopager automatically achieves comparable performance to exhaustive tuning with 10 x less tuning overhead. and 16.3 x and 1.52 x speedup over UMap with default parameters and UMap with page-size only tuning, respectively.
‘Beating the news’ with EMBERS: Forecasting Civil Unrest using Open Source Indicators
Ramakrishnan, Naren; Butler, Patrick; Self, Nathan; Khandpur, Rupinder P.; Saraf, Parang; Wang, Wei; Cadena, Jose; Vullikanti, Anil Kumar S.; Korkmaz, Gizem; Kuhlman, Christopher J.; Marathe, Achla; Zhao, Liang; Ting, Hua; Huang, Bert; Srinivasan, Aravind; Trinh, Khoa; Getoor, Lise; Katz, Graham; Doyle, Andy; Ackermann, Chris; Zavorin, Ilya; Ford, Jim; Summers, Kristen; Fayed, Youssef; Arredondo, Jaime; Gupta, Dipak; Mares, David; Muthia, Sathappan; Chen, Feng; Lu, Chang-Tien (2014)
We describe the design, implementation, and evaluation of EMBERS, an automated, 24x7 continuous system for forecasting civil unrest across 10 countries of Latin America using open source indicators such as tweets, news sources, blogs, economic indicators, and other data sources. Unlike retrospective studies, EMBERS has been making forecasts into the future since Nov 2012 which have been (and continue to be) evaluated by an independent T&E team (MITRE). Of note, EMBERS has successfully forecast the uptick and downtick of incidents during the June 2013 protests in Brazil. We outline the system architecture of EMBERS, individual models that leverage specific data sources, and a fusion and suppression engine that supports trading off specific evaluation criteria. EMBERS also provides an audit trail interface that enables the investigation of why specific predictions were made along with the data utilized for forecasting. Through numerous evaluations, we demonstrate the superiority of EMBERS over baserate methods and its capability to forecast significant societal happenings.

Browse

Browsing College of Engineering (COE) by Content Type "Conference proceeding"

Results Per Page

Sort Options