VTechWorks
VTechWorks provides global access to Virginia Tech scholarship, including journal articles, books, theses, dissertations, conference papers, slide presentations, technical reports, working papers, administrative documents, videos, images, and more by faculty, students, and staff. Faculty can deposit items to VTechWorks from Elements, including journal articles covered by the University open access policy. Email vtechworks@vt.edu for help.
Open Access Policy
Virginia Tech's open access policy enables researchers to deposit the accepted version of scholarly articles with no embargo.
Theses and Dissertations
Virginia Tech was first in the world to require ETDs in 1997, and continues to add scans of older theses and dissertations.
Open Textbooks
More than 50 freely available and openly licensed textbooks are among our most downloaded items.
Communities in VTechWorks
Select a community to browse its collections.
Recent Submissions
Human Auditory Perception and Performance Across Air- and Bone-Conduction Listening Pathways
Currie, Levern Queen (Virginia Tech, 2026-09-09)
Auditory displays support communication, monitoring, and decision-making when visual attention is limited or divided. Most systems use air-conducted (AC) presentation, whereas bone-conducted (BC) presentation delivers vibration through the skull and surrounding tissues while leaving the ear canal open. This open-ear configuration may be useful when users must monitor environmental sound, wear hearing protection, or receive device-mediated audio from more than one source. Three linked gaps remain: performance through AC alone, performance through BC alone, and, most importantly, whether simultaneous AC+BC presentation provides additional usable auditory capacity.
This dissertation addressed those gaps in three within-subjects experiments with untrained participants. Study 1 tested whether participants could identify the presentation mode (AC or BC) and channel (left or right) of tones and noises after subjective reference-tone loudness balancing. Study 2 tested whether simultaneous AC+BC presentation increased the number of signals or spatial locations participants could enumerate and discriminate. Study 3 tested whether AC and BC alerts supported identification and response performance during visual-motor multitasking. Study 1 showed that attribution depended jointly on mode and signal type: BC produced higher accuracy at 250 Hz, whereas AC produced higher accuracy for most mid- to high-frequency tones and white noise. Study 2 showed selected BC advantages for tone-count and spatial-count judgments, but simultaneous AC+BC presentation did not produce a statistically reliable increase in enumeration capacity or spatial discrimination. Study 3 showed higher overall alert-identification accuracy for AC, with mode-specific differences concentrated in broadband alerts. Whether a hybrid AC/BC auditory display is beneficial depends on how information is encoded, what participants must discriminate, and the demands of the concurrent task.
Overall, hybrid AC/BC displays should be applied selectively. The presentation mode, signal characteristics, and task requirements must be matched rather than assuming that BC or simultaneous AC+BC presentation will improve performance. These results provide design guidance for operational, assistive, and multitask auditory displays.
Patterns of Understory Evergreen Shrubs in Two Wilderness Areas of George Washington-Jefferson National Forest using Remote Sensing and Terrain Analysis
Koirala, Pratirakshya (Virginia Tech, 2026-09-09)
This study evaluated the distribution and environmental controls of understory evergreen shrub communities within two federally designated wilderness areas of southwest Virginia at Mountain Lake and Peters Mountain, where human disturbances and access are limited. We focused primarily on the dominant understory evergreen shrubs Rhododendron maximum and Kalmia latifolia. Remotely sensed datasets, topographic variables, and ecological analyses were integrated to investigate relationships between shrub occurrence, topographic gradients, and canopy structure. Landsat imagery and VBMP orthoimages were combined with LiDAR-supported structural information to characterize understory vegetation patterns across heterogeneous forest environments. Spatial analyses demonstrated that evergreen shrub communities were preferentially associated with mesic topographic settings, riparian corridors, north-facing slopes, and areas exhibiting reduced solar exposure and greater moisture retention. The study identified limitations in conventional optical remote sensing approaches for detecting sub-canopy vegetation under dense overstory conditions and offered a new approach to overcome this challenge. The integration of terrain metrics and multi-source geospatial datasets improved shrub mapping accuracy and strengthened interpretation of understory spatial variability. Ecologically, dense shrub layers contribute to reduced tree seedling recruitment, altered nutrient cycling, and shifts in understory biodiversity. These findings emphasize the significance of monitoring understory vegetation for forest management and conservation frameworks in Appalachian ecosystems.
Design and Evaluation of a Memory-Centric Binary Hyperdimensional Computing Architecture with 1T3R Memristor-Based Nonvolatile Storage
Hou, Zeyuan (Virginia Tech, 2026-09-09)
Hyperdimensional computing (HDC) represents information with high-dimensional distributed hypervectors and performs classification using simple binary operations such as binding, bundling, and associative search. While these operations are hardware-friendly, the large hypervector dimension creates a substantial persistent-memory and data-movement burden because position/identifier (ID), level, and trained class hypervectors all scale linearly with dimension D.
This thesis develops and evaluates a memory-centric Binary HDC architecture for the Modified National Institute of Standards and Technology (MNIST) handwritten-digit dataset. Persistent binary hypervectors are mapped to nonvolatile one-transistor/three-resistor (1T3R) memristor storage, while nearby complementary metal-oxide-semiconductor (CMOS) circuitry performs exclusive OR (XOR) binding, bundling, majority thresholding, Hamming-distance accumulation, and class selection. A controlled design-space study isolates the effects of input resolution, model adjustment, quantization, and hypervector dimension. The selected Balanced D = 5,000 configuration achieves 86.87% validation accuracy and 87.74% accuracy on the official MNIST test set after model freeze.
The hardware mapping treats the three memristors in each 1T3R row as three independent HDC dimensions rather than one signed 3-bit coefficient. For the frozen full-resolution, M = 32 configuration, persistent storage scales as 826D logical bits, with the ID bank accounting for approximately 94.9% of capacity. Under a direct-access baseline without caching, one query requires 1,578D logical persistent-bit reads. Measured level usage also reveals an idealized within-query Level-hypervector reuse opportunity that could reduce persistent read traffic by approximately 47.95% with suitable local buffering.
Circuit evidence is obtained from the existing SkyWater 130 nm 4 × 3 schematic proof of concept. Experiment 8 (EXP8) verifies all eight Row 1 three-bit states and all 24 static bit decisions, while selective SET of G2, G1, and G0 preserves all observed neighboring bit states. The high-resistance-state (HRS) and low-resistance-state (LRS) current populations remain clearly separated, with a worst-case LRS/HRS ratio of 140.94×. Robust voltage-domain sensing, multi-row disturb, independent RESET, process, voltage, and temperature (PVT) variation and Monte Carlo characterization, and complete-system area, energy, and latency evaluation remain outside the completed scope.
WIP: False-Data Injection Against Physics-Constrained Neural Networks for Trajectory Planning in Autonomous Racing
Yang, Chun-Tao (Virginia Tech, 2026-09-09)
Prediction of competitors' motion is critical for high-speed autonomous racing, yet its robustness under adversarial manipulation remains unclear. We analyze black-box false- data-injection (FDI) attacks that corrupt perceived motion histories, inducing ego mispre diction and enabling adversarial overtaking. Our results reveal that high-speed trajectory prediction systems remain vulnerable, exposing safety risks in multi-agent autonomous driving
Analyzing Surface Urban Heat Island Trends and Regional Variability Across 150 U.S. Cities (2000–2024) Using MODIS Data and a Reproducible Python Application
Gautam, Suyog (Virginia Tech, 2026-09-08)
Urban Heat Islands (UHIs), or Surface Urban Heat Islands (SUHIs), occur when urban surfaces register measurably higher temperatures than surrounding non-urban land, a consequence of replacing natural vegetation with impervious materials that absorb and re-emit solar radiation. As more than 80% of the U.S. population now lives in urban areas, the thermal consequences of continued urbanization carry growing implications for public health, energy demand, and climate adaptation planning. This study analyzes SUHI dynamics across 150 U.S. cities from 2000 to 2024 using MODIS land surface temperature data, an automated Google Earth Engine (GEE) data extraction application, and multi-year trend analysis. Three questions are addressed: (i) whether an open-source, GEE-based Python application can reliably automate consistent SUHI extraction for any U.S. city; (ii) how regional SUHI trends and temporal trajectories differ across five U.S. geographic regions and three city-size classes over the 25-year study period; and (iii) how the Diurnal Asymmetry Index (DAI) identifies whether cities have stronger daytime or nighttime SUHI and how these differences are related to regional hydroclimate conditions. Our results show that urban heat trends are not the same across the United States. Some regions experience much stronger SUHI changes than others. We also found that after around 2018, the strengthening of SUHI generally slowed in many cities, although the amount of slowdown varied by location.
Another key result is that daytime and nighttime SUHI behavior separates into two distinct patterns, and those patterns align closely with climate zones defined by the Köppen–Geiger classification. Finally, the automated system successfully produced standardized and reproducible annual metrics for all 150 cities, showing that large-scale SUHI monitoring can be done in an open and accessible way.