Virginia Tech - Wake Forest University School of Biomedical Engineering & Sciences
VTech Works (Virginia Tech)Not a member yet
122049 research outputs found
Sort by
Physiological Sensing for Driver State Monitoring: Technology Scan and Pilot Evaluation
No full textAs advanced driver assistance systems and automated vehicle technologies evolve, the ability to monitor and assess driver readiness remains critical. In support of that need, this report describes several efforts to evaluate the feasibility and reliability of capturing physiological signals during real-world driving. The goal was to examine whether signals from respiration, cardiac activity, and brain function, captured via wearable and non-contact sensors, could complement existing driver monitoring methods and provide useful input for future in-vehicle systems. A review of the literature supporting respiration, cardiac, and brain activity as relevant domains for driver state monitoring was the initial step in this process. These physiological channels are discussed as potentially useful complements to traditional measures like eye behavior, especially given their links to autonomic and cognitive changes under various degraded states. The literature review was complemented by a technology scan of commercial and research-grade devices capable of measuring these signals in mobile contexts. A structured protocol for exercising an initial subset of these systems during impaired driving in a closed-course environment was also developed. This protocol was then executed in a pilot test with five participants, who completed standardized baseline and post-alcohol drives while instrumented with electrocardiogram (ECG), respiration, and electroencephalography (EEG) sensors in addition to sensors capturing driving performance and glance behavior. In this pilot study, alcohol was used as a convenient physiological stressor given its well-understood dosage effects on driving and physiology. Results indicated that alcohol consumption consistently altered some behavioral physiological signals across all three domains. Physiological responses were more robust and consistent than the observable driving metrics, potentially highlighting the complementary value of these signals. The small sample size, however, resulted in a lack of power to detect statistical significance between sober and impaired driving for most metrics. Respiration and ECG signals were captured with high reliability, while EEG results provided informative patterns but suffered from variable signal quality and motion artifacts. These findings support the initial viability of cardiac and respiratory sensing in mobile in-vehicle settings and highlight practical limitations that must be addressed for EEG. Altogether, the effort demonstrates that it is technically feasible to capture and interpret physiological signals in a real-world driving context using wearable and embedded sensors. While further validation is needed, the results provide a foundation for integrating such signals into future driver state monitoring systems, not as standalone indicators, but as part of a multimodal approach that reflects the complexity of driver physiology and behavior
Evaluation of Driver Response to Arrow Boards of Varying Patterns
No full textTo investigate potential impacts of various arrow board display patterns on traffic behavior, a semi-naturalistic study was conducted within a live work zone in Salem, Virginia. Specifically, the arrow board displays examined in the study were flashing arrow, sequential arrow, and sequential chevron patterns. The data collection site for the study was a work zone with a semi-permanent lane closure on Wildwood Road in Salem, Virginia, underneath the I-81 overpass at Exit 137. An arrow board was placed within a taper of traffic barrels in the left northbound lane, approximately 220 ft upstream of the work area. The taper redirected northbound traffic on Wildwood Road into the right lane while the arrow board displayed an appropriate pattern indicating the merge direction to oncoming vehicles. A standard 25-light arrow board was acquired and equipped with a data collection system that included a camera and radar sensor. The arrow board was then positioned into the existing work zone. The data collection system was programmed to automatically record video and radar data twice per day: 10:00 a.m. to 11:00 a.m. (i.e., morning) and 4:00 p.m. to 5:00 p.m. (i.e., afternoon). A different display pattern was used each week, with data collection occurring for five consecutive days (Sunday through Thursday) per display pattern. A supplemental data collection system was also placed upstream of the work zone to capture baseline speed data for traffic prior to when the arrow board was visible. Video from the arrow board was reduced to determine the number of vehicles in each lane as they approached the arrow board, as well as the distance at which vehicles merged from the closed lane into the open lane relative to the arrow board. Radar data from the arrow board was also reduced to isolate speed measurements to only vehicles traveling on Wildwood Road that ultimately passed the arrow board leading into the work zone. Vehicles that turned or entered from an intersecting street were excluded from analysis. Results of the study showed that time of day (i.e., morning vs. afternoon) had a significant effect on measures of traffic volume and lane occupancy rate. Traffic volume was found to be significantly higher during the afternoon (105 vehicles per 10 min) than in the morning (64 vehicles per 10 min). In contrast, the lane occupancy rate (i.e., the rate of vehicles traveling in the closed lane) was significantly higher in the morning (7.6%) than in the afternoon (5.1%). The arrow board display pattern was not significant for either traffic volume or lane occupancy rate. Neither the display pattern nor time of day had a significant effect on the distance at which vehicles changed from the closed lane into the open lane. However, the interaction between the two was approaching statistical significance (p = 0.07). Mean lane change distances ranged between 296 ft and 343 ft across all display patterns and times of day. However, an evaluation of the vehicle merge rate by distance revealed interesting differences among the display patterns. During the morning, the flashing arrow pattern consistently had the lowest proportion of vehicles using the closed lane at all distances, while the sequential chevrons had the highest proportion. During the afternoon, the sequential arrow had a consistently lower proportion of vehicles remaining in the closed lane across all distances, while the flashing arrow and sequential chevrons performed worse but similarly to each other
From Compliance to Culture: Ethics in Agricultural Education Research
No full textEthics in agricultural and life science research is often treated as a matter of compliance, yet the integrity of our work—and public trust in it—depends on everyday decisions made in labs, classrooms, and communities. This interactive session adapts Virginia Tech’s Innovative Research and Ethical Impact (IREI) model to the context of the American Association for Agricultural Education (AAAE), positioning Southern Region scholars as catalysts for ethical cultures in academia.
Participants will (1) explore how social norms and informal mentoring shape “how we do things” in research groups and departments, (2) identify ethical dimensions within their own current or planned projects across the research pipeline—from formulating questions to disseminating findings, and (3) practice strategies for communicating and addressing ethical concerns in ways that align with AAAE Research Values and AFNR priorities such as environmental health, diversity and inclusion, youth development, and safety.
Using brief case scenarios and peer discussion, we will foreground questions of who is affected, how, and with what scope and severity, emphasizing the human consequences of decisions in agricultural and related social science research. The session is designed for faculty, graduate students, and academic leaders who want practical, discipline-relevant tools for moving beyond one-off responsible conduct of research trainings toward a sustained community of practice around ethics in academia. Participants will leave with concrete conversation prompts, reflective questions for research teams and classrooms, and an action plan for modeling ethical research and teaching practices within their own spheres of influence
Angewandte Chemie International Edition
No full textMolecular definition is usually regarded as a prerequisite to achieve protein recognition and functional modulation, particularly for macromolecular interactions. Herein, we report that polymers with specific combinations of monomers arranged into random sequences [random hetero oligomers (RHOs)] can selectively bind to a model protein. Using green fluorescent protein (GFP) as a target, polyacrylates were developed that bound with nanomolar affinity and enhanced fluorescence by >100%. Purification of the polymerization product revealed subpopulations of compositions with distinct affinities and selectivity for GFP over a competing protein. Experimental and computational binding analyses confirmed that there are distinct RHO–GFP interactions, which are influenced by RHO chemical composition. These findings show that sequence-defined structures are not a prerequisite for selective protein recognition. Synthetic polymers can instead serve as scalable, tunable platforms for molecular recognition—representing a significant leap towards next-generation sensing, therapeutic, responsive, and catalytic materials in domains previously dominated by biologics or complex peptide scaffolds.Published versio
Shorebirds in a changing landscape: Assessing the top-down and bottom-up drivers of American oystercatcher (<i>Haematopus palliatus</i>) reproductive success in the Virginia barrier islands
No full textUnderstanding the factors that affect population growth is a vital component of the conservation of imperiled species. On the United States (U.S.) Atlantic and Gulf coasts, the American oystercatcher (Haematopus palliatus) is largely managed within a conservation framework that aims to identify and manipulate factors that limit reproductive success. Coordinated research and monitoring across the species' range have identified several key threats to the survival of nests and chicks, including predation, habitat loss and degradation, and human disturbance. However, the relative importance of those threats can change, and new threats may emerge, as coastal ecosystems dynamically respond to global climate change. In the Virginia barrier islands, where American oystercatcher reproductive success has been historically high due to the widespread availability of undeveloped coastal habitat, declines in the average annual productivity rate of American oystercatchers since 2016 may signal evolving threats to reproduction.
To provide insight into changing threats to American oystercatcher productivity in Virginia, and to inform adaptive management, we investigated the drivers of American oystercatcher reproductive success on the Virginia barrier islands. We considered three components of American oystercatcher conservation and management: (1) interactions between American oystercatchers, their conspecifics, and a complex predator community; (2) altered availability of nesting habitat due to climate-driven changes within barrier island ecosystem; and finally, (3) community support for widespread protection of the barrier island landscape and participation in conservation initiatives designed to protect breeding American oystercatchers, such as adhering to island access restrictions and recreation guidelines. We first investigated nest and chick survival on Metompkin Island and Fisherman Island—two islands that represent a range of geomorphological and ecological conditions present within the Virginia barrier island system—in 2021–2023 using a combination of field-based surveys and automated radio telemetry methods. Next, we used remotely sensed data to explore how climate-driven ecosystem state change on the Virginia barrier islands may be driving changes in the abundance and distribution of American oystercatcher nesting habitat across the barrier island system from 2004–2021. Finally, we evaluated the short-term outcomes of a field-based high school environmental education program using pre- and post-program surveys of students to assess how environmental education may promote changes in knowledge and public attitudes related to shorebird conservation initiatives in Virginia.
Through routine surveys of nests and radio-marked chicks, we found that the cumulative probability of a nest surviving the incubation stage was greater than 90%, while the cumulative probability of a chick surviving from hatch to fledging was approximately 51%. Thus, chick survival, rather than nest survival, may be limiting reproductive success in the Virginia barrier islands. Flooding was a primary cause of nest loss (32% of failed nests), and storm-driven overwash in late April and early May forced peak nesting to shift later in the breeding season as breeding pairs renested, particularly in 2022. Despite ongoing management of mammalian predators through lethal removal, reproductive success still appears to be limited through the top-down effects of predation on both stages of American oystercatcher reproduction, as both nests (22% of failed nests) and chicks (46.4% of marked chicks) were lost from non-mammalian predators, including raptors (e.g., peregrine falcons [Falco peregrinus] and great horned owls [Bubo virginianus]) and Atlantic ghost crabs (Ocypode quadrata).
Remotely sensed data provided an opportunity to investigate factors for which we lacked data collected in situ. Using aerial orthoimagery of the barrier islands collected by the National Agriculture Imagery Program (U.S. Department of Agriculture), we found that habitat abundance increased by 228.26 ha from 2004–2016, and then decreased by 124.38 ha from 2016–2021. Across the 2004-2021 timespan, abundance of American oystercatcher nesting habitat (defined as sites where the relative probability of nesting is ≥ 90%) was both temporally and spatially variable within the Virginia barrier island systems, with changes on only a few islands (e.g., Metompkin Island, Cedar Island, and Cobb Island) driving most of observed system-wide trends. We suggest that temporal variability may be a result of patterns of storminess along the Virginia coast 2004–2021, while spatial variability is likely due to localized differences in the geomorphological and ecological processes that control habitat availability, including sediment dynamics (e.g., sand erosion versus accretion), dune-building processes, and vegetative succession.
Finally, we used an evaluation of The Nature Conservancy's field-based environmental education program for high school students as a case study to demonstrate the value of incorporating environmental education into a multidisciplinary conservation program. Using surveys that assessed components of environmental literacy—including knowledge of ecological concepts, feelings of connection and stewardship toward the local ecosystem, and behavioral intentions—we examined short-term educational outcomes for students who participated in a formal field trip to one of the Virginia barrier islands. By comparing responses on surveys delivered immediately before and immediately after the field trip, we found that participation in the field trip increased students' knowledge about barrier island ecosystems and sense of place attachment. Additionally, students self-reported positive environmental literacy outcomes on the post-trip assessment, including increased interest in scientific learning, increased likelihood of participating in stewardship behaviors, and increased intentions to participate in conservation behaviors.
Given the pace at which ecosystems are changing in response to climate change and anthropogenic drivers, it is expected that the factors limiting American oystercatcher population growth will also continuously change. To protect against population declines, shorebird managers will need to continuously monitor for change. They will also need to consider the use of creative management approaches, including ecosystem-based management to address complex predation and climate threats, and environmental education and outreach programs to promote broad community support for conservation initiatives.Doctor of PhilosophyThe American oystercatcher is a large shorebird that nests along the United States Atlantic and Gulf coasts, from Maine to Texas. American oystercatchers have faced population declines due to threats from predators, changes in the amount and quality of their habitat, and conflicts with human communities that use similar coastal areas for recreation and commercial purposes. While efforts have been made to address these threats, new challenges to American oystercatcher conservation are emerging due to global climate change.
The Virginia barrier islands are a mostly undeveloped coastal landscape located on the Atlantic coast of the Eastern Shore of Virginia, which includes Accomack County and Northampton County. The island system stretches for approximately 110 km (68 miles), providing important habitat for breeding American oystercatchers. Virginia has historically supported large numbers of breeding American oystercatchers, as well as high rates of reproductive success. However, since 2016, average yearly rates of American oystercatcher reproductive success in the Virginia barrier islands have been lower-than-expected despite continued management actions taken by The Nature Conservancy, Virginia Department of Wildlife Resources, United States Fish and Wildlife Service, and other state and federal government agencies. Biologists suspect that changes to American oystercatcher populations in Virginia may be, in part, a result of sea-level rise, different patterns of storm intensity and frequency, and warming temperatures.
American oystercatcher reproduction occurs within two stages: (1) a nesting stage and (2) a chick-raising stage. Combined, these stages determine productivity, or the number of chicks that each breeding pair contributes to the population annually. We investigated the factors that contribute to American oystercatcher productivity in Virginia. First, we monitored nest and chick survival on two islands, Metompkin Island and Fisherman Island, in 2021–2023, so that we could calculate the probability of nests and chicks surviving each stage of reproduction and identify causes of nest loss or chick death. Next, we used aerial imagery collected by aircraft from 2004–2021 to answer the question: 'has the amount and location of nesting habitat changed for American oystercatchers in Virginia over the last eighteen years?' Finally, we examined one of The Nature Conservancy's school field trip programs on the Eastern Shore of Virginia to understand how much students learned during the trip, as well as how their feelings and attitudes about the Virginia barrier islands changed after participating in the trip.
From monitoring American oystercatcher nests and chicks on Metompkin Island and Fisherman Island, we found that the probability of chicks surviving was lower than the probability of a nest surviving. Flooding from storms was an important cause of nest loss. Predation by other species of birds (such as peregrine falcons and great horned owls) and Atlantic ghost crabs was an important cause of both nest loss and chick deaths.
We found that the amount and locations of American oystercatcher nesting habitat changed over time within the Virginia barrier islands. Additionally, each island had different patterns of change. For example, some islands, such as Metompkin Island, Cedar Island, and Cobb Island, gained a substantial amount of habitat during periods within 2004–2021, while other islands, such as Hog Island and Parramore Island, did not. Overall, the amount of American oystercatcher nesting habitat increased within the system by 228.26 ha (about 564 acres) from 2004–2016, and then decreased by 124.38 ha (about 307 acres) from 2016–2021.
Finally, we used an evaluation of The Nature Conservancy's high school field trip program as an example of how environmental education programs can have a positive impact on conservation by increasing public knowledge about conservation issues and helping people feel attached to important natural resources. We delivered surveys to students before and after their field trip and asked them to answer questions related to their experience, what they learned, and how much they agreed with pre-written statements of feelings regarding the barrier islands and natural environment. When we compared students' responses from before and after the trip, we found that students knew more about barrier islands and felt a greater sense of attachment towards the barrier island system after returning from the field trip. Additionally, students reported that they were more interested in learning science and more interested in participating in conservation efforts after going on the field trip.
As Virginia's barrier islands and its shorebirds continue to change due to changing climate, land managers such as The Nature Conservancy will need to continue to monitor and manage for changing predator populations, potentially restore habitats so that they are resilient to storms and rising sea levels, and continue to work closely with students and adults on the Eastern Shore of Virginia to create and enact the most sustainable conservation actions.
10C<small>ACHE</small>: Heterogeneous Resource-Aware Tensor Caching and Migration for LLM Training
No full textTraining large language models (LLMs) in the cloud faces growing memory bottlenecks due to the limited capacity and high cost of GPUs. While GPU memory offloading to CPU and NVMe has made large-scale training more feasible, existing approaches suffer from high tensor migration latency and suboptimal device memory utilization, ultimately increasing training time and cloud costs. To address these challenges, we present 10Cache, a resource-aware tensor caching and migration system that accelerates LLM training by intelligently coordinating memory usage across GPU, CPU, and NVMe tiers. 10Cache profiles tensor execution order to construct prefetch policies, allocates memory buffers in pinned memory based on tensor size distributions, and reuses memory buffers to minimize allocation overhead.
Designed for cloud-scale deployments, 10Cache improves memory efficiency and reduces reliance on high-end GPUs. Across diverse LLM workloads, it achieves up to 2× speedup in training time, improves GPU cache hit rate by up to 86.6×, and increases CPU/GPU memory utilization by up to 2.15× and 1.33×, respectively, compared to state-of-the-art offloading methods. These results demonstrate that 10Cache is a practical and scalable solution for optimizing LLM training throughput and resource efficiency in cloud environments.Published versio
Impact of Prescribed Fire on the Terrestrial Orchid, <i>Isotria medeoloides</i> in Virginia's Blue Ridge Mountains
No full textSmall whorled pogonia (Isotria medeoloides [Pursh] Raf.) is a terrestrial orchid with imperiled G2S2 conservation status. It is native to Appalachia from Georgia to Quebec. This orchid is one of many species whose habitat is being lost to degradation. This mycoheterotrophic orchid is seasonally dormant, but can also remain fully dormant with no above-ground tissue for multiple years. Conservation efforts over the last forty years have provided a management framework that suggests fire may be useful to recovering this flower and restoring its habitat for the benefit of the species that share it. Our study site is one of these habitats where small whorled pogonia grows in a region with a historical regime of frequent fire. The study site is on the western slopes of Virginia's Blue Ridge and is one of many natural area preserves that is managed by the Virginia Department of Conservation and Recreation. This preserve is undergoing ecological restoration with an objective of restoring historical open-canopy oak-hickory woodlands. Open oak woodlands were typical of the historical fire regime in the Shenandoah Valley (Lafon et al., 2017). As a threatened species, small whorled pogonia is protected under the Endangered Species Act of 1973. This requires that publicly funded species management must mitigate harm. Before fire can be used in populations of small whorled pogonia on public lands, its impacts on the orchid and its habitat must be examined. The DCR Natural Heritage Program, in collaboration with Virginia Tech, put together a plan to study the impacts of prescribed fire on the rare orchid population. This observational study gathered multivariate environmental data during three growing seasons leading up to the fire and two seasons following the 2024 fire. The population was geographically divided, and one subpopulation was excluded from prescribed burning. The smaller subpopulation underwent an operational low-intensity prescribed fire in March 2024. Destructive sampling was used to assess fuel loading before and after the fire. Canopy imaging and biodiversity plots were used to assess ecosystem response to fire. Generalized mixed effects models, multilevel pairwise comparisons, and local spatial analysis including Getis Ord Gi and Lee's L were used to explore the individual, population, and community level response to fire. Researchers examined the effects of fire on small whorled pogonia stem counts and life stages and the effects of fire on plot-level alpha diversity and forest floor cover and light transmission. A population differential was calculated by subtracting the burned population stem count from that of the unburned. Results showed that the population differential increased by 17% in spring 2024 (two months after fire) but in 2025, following a steep decline in the unburned population stem count, the differential decreased by 75%. The population persisted after fire with no significant effect on seedling establishment. Dormancy was not measured separately from death. Likelihood of dormancy increased for both populations since 2022, with greater amplitude for the burned population in 2024 and for the unburned population in 2025. Flowering was not affected, but fruiting rate increased in 2024 for the burned population. Low intensity prescribed fire had little effect on forest floor light transmission. Forest floor cover was reduced and favored graminoid and forb cover over woody cover. Woody cover by Pinus strobus saplings was reduced in the shrub layer. Overall, the single observed fire event had short-term effects that reduced small whorled pogonia stem count and did not increase forest-floor richness or light transmission. Further observation of this population will be necessary to disentangle plant death and plant dormancy as fire responses within this population and to assess post-fire population viability. This study describes the isolated effects of a single fire on one subpopulation of small whorled pogonia and its results are not representative of a species-wide fire response. Further study should examine long-term and repeated use of frequent fire across multiple small whorled pogonia populations.Master of Science in Life SciencesSmall whorled pogonia (Isotria medeoloides [Pursh] Raf.) is a G2S2 federally threatened terrestrial orchid. It is native to Appalachia from Georgia to Quebec. This orchid is one of many species whose habitat is being lost to degradation. Fire could help this endangered orchid and its habitat to recover. This study focuses on one population of small whorled pogonia in the Shenandoah Valley watershed on a Virginia Department of Conservation and Recreation Natural Area Preserve. This preserve is undergoing ecological restoration with the objective of restoring the historical open-canopy oak-hickory woodlands. To determine the effects of fire on small whorled pogonia, this observational study compared the population dynamics of two subpopulations of the species at the Natural Area Preserve: one site was burned twice and the other was excluded from fire. Researchers examined the effects of fire on stem counts, fecundity, canopy openness, and site biodiversity. Results showed that the burned population declined in size relative to the unburned population in spring 2024 (two months after fire) but in 2025, following a steep decline in the unburned population stem count, the relative size of the burned population increased. The population persisted after fire with no significant effect on flowering or seedling establishment. Low intensity prescribed fire had little effect on canopy openness and site biodiversity
MedEdPORTAL : the journal of teaching and learning resources
No full textIntroduction: Trauma affects 90% of individuals and has profound impacts on health, making it essential for medical trainees to recognize its effects. Trauma-informed care (TIC) offers a framework for developing these skills. Despite its importance, no TIC curriculum integrates community feedback into its design. To address this gap, we developed a 4-hour TIC curriculum that incorporates community insight, clinical expertise, and practical communication training.
Methods: The curriculum design followed community-based participatory research principles, engaging community members as contributors. The training included a dynamic combination of didactic lectures, video demonstrations, small-group role-play, and an OSCE, supported by a novel TIC toolkit. Community partners were trained as standardized patients (SPs). We assessed student outcomes through pre- and postsession surveys, employing 5-point Likert scales and open-ended responses. Additionally, a custom assessment tool was developed to evaluate OSCE performance, with SPs providing structured feedback.
Results: Thirty-four third-year medical students participated, with 100% survey completion. Quantitative analysis revealed significant increases in students' understanding of TIC principles and confidence in applying them from pre- to postsession (p < .05 for all metrics). Students demonstrated strong performance on the OSCE, achieving a mean OSCE performance score of 31.4/38 (or overall score of 82.6%). SP feedback highlighted the students' ability to engage empathetically and effectively in trauma-sensitive encounters.
Discussion: This novel TIC curriculum on women's health demonstrates a successful, scalable model for integrating TIC training into medical education. By embedding community voices and combining evidence-based principles with experiential learning, this program addresses educational gaps in TIC medical education.Published versio
Artificial General Intelligence (AGI)-Native Wireless Systems: Digital Twins and World Models for Beyond 6G Networks
No full textBuilding next-generation wireless systems that can reliably support physical artificial intelligence (AI) agents (e.g., robots, autonomous vehicles, etc.) requires advanced levels of intelligence beyond today's state-of-art. On the one hand, the 6G vision of AI-native networks typically relies on standard AI methods (e.g., neural networks) that perform poorly in non-stationary, real-world environments. On the other hand, physical AI agents still lack the capability to generalize and adapt in unforeseen scenarios that appear in real-world settings. As a result, today's AI-native wireless systems and, in turn, their physical AI agents remain far from being autonomous and fall short in terms of quality-of-service. To address this limitation, this dissertation aims to revisit and redefine the concept of AI-native wireless systems, equipping them with common sense capabilities necessary to transform them into artificial general intelligence (AGI)-native systems. This transformation promises to revolutionize wireless systems by enabling unprecedented levels of cognitive wireless intelligence. Notably, such intelligence provides networks with the reasoning, planning, and complex inference needed to operate in dynamic, real-world environments. This envisioned new generation of networks is driven by a cognitive brain architecture that is founded on three components: A perception module, a world model, and an action-planning component. Towards realizing these components, first, we show how the perception module can be built through abstracting real-world elements into generalizable representations. Then, these representations are used to form a world model, founded on principles of causality and hyper-dimensional computing. Subsequently, we design intent-driven and objective-driven planning methods that can maneuver the network to take its actions. Central to this architecture, world models offer a structured approach for mirroring the physical world into a digital counterpart over the network. Cheif among these counterparts are the digital twins (DTs) of physical AI agents. With this interconnection to world models, DTs offer a gateway to instill common sense from the network directly into these agents. Nevertheless, enabling such solution requires addressing novel wireless challenges. Chief among those challenges is preserving the synchronization of DTs and world models with the physical world. To address this challenge, we propose a rigorous decentralized framework that decomposes these world models and their DTs over the network edge. In particular, we pose an optimization problem that aims to minimize the synchronization delays of smaller-scale world models and their associated DTs at the edge, while ensuring their interoperability in terms of association and resource allocation. To solve this problem, we propose an optimal transport theory algorithm that ensures the optimal average synchronization time of the world models, while satisfying the synchronization intensity requirements of the DTs. Results show that synchronization delays can be reduced up to 25 % in comparison to the standard signal-to noise ratio (SNR) association benchmark. Accordingly, the decentralized DTs and world models are then leveraged to drive reasoning back into the physical AI agents in the real world. This reasoning allows the physical AI agents to generalize when facing unforeseen scenarios, thereby enabling a revolutionary test-time scaling law for physical AI agents. In particular, this novel scaling law builds on the first principle of active inference and extends action selection to incorporate inference-driven reasoning that scales the feed-forward policy in unforeseen scenarios. Nevertheless, this decision-making process is formulated as a partially observable Markov decision process (POMDP) that renders an intractable inference problem. To obtain a tractable solution for this POMDP, this problem is solved through a variational Bayesian approach that unifies perception, planning, action, and learning under the minimization of variational and expected free energy. Results showcase how the proposed framework yields AI agents that can act, reason, learn, and maintain generalizable performance in dynamic environments. As continual learning (CL) abilities emerge while updating the DT models with these unforeseen scenarios, we further design a deep CL solution to enable synchronized model updates for physical AI agents. In particular, we propose a novel CL solution that preserves the accuracy and synchronization of the evolving DTs at the edge. To limit the de-synchronization gap arising during the DT model update, this update process is posed a dual objective optimization problem whose goal is to jointly minimize the loss function over all encountered historical episodes and the corresponding de-synchronization time. As the de-synchronization time continues to increase over sequential historical episodes, an elastic weight consolidation (EWC) technique that continually regularizes the DT history is proposed to limit de-synchronization time. Furthermore, to address the plasticity-stability tradeoff accompanying the progressive growth of the EWC regularization terms, a modified EWC method that considers fair execution between the historical episodes of the DTs is adopted. Simulation results show that the proposed solution can achieve an accuracy of 90% while guaranteeing a minimal de-synchronization time. Therefore, this dissertation is expected to shape the future of DTs and world models as enablers of AGI over future networks. Ultimately, this dissertation serves as a blueprint to drive the next generation of wireless networks and its autonomous physical AI agents in the 6G and beyond era.Doctor of PhilosophyNext-generation telecommunication networks (e.g., 6G cellular systems) and artificial intelligence (AI) systems are evolving towards agentic frameworks that autonomously interact with the physical world. While the current generation of AI has shown tremendous impact in fields like language, mathematics, and coding, real-world autonomous agents (e.g., robots, autonomous vehicles, etc.) and telecommunication systems still fall short in showing similar competing performance. To address this limitation, a shift towards AI architectures that support reasoning, planning, and complex inference to deal with the dynamic nature of real-world environments is necessary. This dissertation explores how the intersection of digital twins (DTs) and world models plays a role in enabling these new architectures. In essence, a world model is an internal representation of an environment that encodes its states, dynamics, and uncertainty, enabling an agent to predict outcomes, plan actions, and reason about future scenarios. Notably, integrating world models into next-generation networks provides a unique opportunity to develop a common sense understanding of "how the world works." This ability is a cornerstone for dealing with the countless, unforeseen scenarios that agents encounter in the real world. Effectively, it enables next-generation networks to generalize beyond their training domain and drive new levels of network intelligence. Nevertheless, to enable world models over the network, telecommunication systems should still acquire core cognitive abilities such as perception, abstraction, and analogy. To provide these missing cognitive abilities and close the loop, we present the first cognitive brain architecture tailored to a telecommunication system. Subsequently, we elucidate the design of the cognitive modules embedded into this brain architecture. Ultimately, this cognitive architecture serves as a foundation for transitioning towards artificial general intelligence (AGI)-native networks in the beyond 6G era. More broadly, this advanced level of intelligence further expands beyond the dimensions of the network to augment its autonomous agents. In particular, DTs mirror the physical states of autonomous agents into these world models. The network thus enables agents to reason about their environment and adapt to unforeseen conditions. To this end, DTs are designed from first principles to enable test-time scaling for these agents. Accordingly, algorithms for decentralized world models and continual learning agents at the network edge are proposed. Overall, this dissertation lays the foundations of DTs and world models that promise to advance the intelligence levels of autonomous agents and telecommunication systems in the beyond 6G era
Lecture Notes on Configuration Aerodynamics
Full text linkPractical applications of aerodynamic theory are critically important for aerodynamic design of aircraft configurations, but are often omitted from aerospace engineering curricula. Lecture Notes on Configuration Aerodynamics is an incredible resource for educating the next generation of aerospace engineering students who aspire to engage in
the aerodynamic design of aircraft configurations. It shows how aerodynamic theory is applied in practice, giving students insight into what a career in aerodynamics entails. This textbook offers a design-oriented perspective of the development and analysis of aircraft aerodynamics.
Based on his academic experience as a professor and his industrial experience at Grumman, Mason presents decades of relevant knowledge and wisdom of a large number of exceptional researchers and practicing engineers. Extensive references throughout the book encourage further study of configuration aerodynamics.
This open textbook encompasses the aerodynamic design of flight vehicles with emphasis on flow fields and configuration concepts. Mason covers methodologies for aerodynamic analysis and design for flows ranging from low speed to high speed and includes case studies of classic configurations.
Are you reviewing or adopting this book for a course?
Please help us understand your use by filling out this form.
How to access this book
The open textbook is freely available online in multiple formats, including: PDF, EPUB, and Pressbooks.
A paperback print version (in color) is available for order here.
This page may include links to supplementary resources. The permanent URL for this page is: https://hdl.handle.net/10919/139932.
The publisher's page for this book is: https://doi.org/10.21061/configurationaerodynamics.
ISBNs
ISBN (PDF): 978-1-962841-52-8
ISBN (Pressbooks): 978-1-962841-53-5
ISBN (EPUB): 978-1-962841-51-1
ISBN (print): 978-1-962841-50-4
Table of contents
1. Introduction to Configuration Aerodynamics
2. Foundations of Fluid Mechanics: Governing Equations
3. Fundamentals of Aerodynamic Drag
4. Configuration Aerodynamic Design: Use of Computational Aerodynamics
5. Subsonic Aerodynamics: Airfoils and Wings
6. Transonic Aerodynamics: Airfoils and Wings
7. High-Lift Aerodynamics
8. High-Angle-of-Attack (High-α) Aerodynamics
9. Supersonic Aerodynamics
10. Hypersonic Aerodynamics
11. End Note
Appendix A: Geometry for Aerodynamicists
Appendix B: Fifteen Minutes of Stealth in Aircraft Design
Appendix C: Government Regulations Affecting Configuration Aerodynamics
Appendix D: Examples of Aerodynamic Design
Appendix E: Software for Aerodynamic Analysis and Aircraft Design
Appendix F: Configuration Aerodynamics Reading List
Appendix G: The Configuration Aerodynamicist’s Bookshelf
About the author and editor
William H. Mason, author
William H. Mason (1947–2019) developed a deep passion for airplanes quite early in his life. Growing up in Southwest Virginia, he spent countless hours building and flying model airplanes as a teenager. When he was an undergraduate student at Virginia Tech, he seized upon opportunities to gain practical experience, working summers at McDonnell Douglas in St. Louis, Missouri, where he was involved with various F-4 aircraft projects, including the swing-wing F-4, and at the Edwards Air Force Base, California, working on US Army Huey Cobra helicopters. In 1974, he began his fifteen-year professional aerospace engineering career with Grumman, where he made valuable contributions to many high-profile projects, such as: (i) the X-29, an experimental aircraft with a forward-swept wing and canard; (ii) the NASA/Grumman Research Fighter Configuration with supercruise and maneuvering capabilities; and (iii) the SC3 Wing Concept, which set a record for low drag at high-lift supersonic performance. From 1989 until his passing in 2019, he was a dedicated educator at Virginia Tech. Right after returning to VT in 1989, he devoted himself to sharing his knowledge and insights with students and colleagues. His legacy lives on with a large number of students who either took the courses he offered in aircraft design, applied computational aerodynamics, and configuration aerodynamics or performed research in aerospace systems design and multidisciplinary optimization. He co-authored Applied Computational Aerodynamics: A Modern Engineering Approach—one of the first textbooks on this topic for undergraduates—published by Cambridge University Press in 2015. He also authored or co-authored more than 100 technical papers and reports. He was a lifelong Hokie, having earned his BS degree in 1971, a MS in 1972, and a PhD in 1975, all in aerospace engineering from Virginia Tech. Mason was an Associate Fellow of the American Institute of Aeronautics and Astronautics (AIAA).
Pradeep Raj, editor
Pradeep Raj is a Collegiate Professor Emeritus at Virginia Tech and spent twelve years serving as a faculty advisor of student capstone aircraft design teams and conducting collaborative research in simulation driven design to enable development of quality affordable aerial vehicles. He joined VT in 2012 after thirty-two years (1979–2011) with Lockheed Martin, a premier aerospace and defense corporation. For the first twenty years there, he held key technical leadership positions and made noteworthy contributions to advancing the effectiveness of computational simulation capabilities for meeting aircraft design needs. For the next twelve years, he held executive leadership and management positions before retiring from the Advanced Development Programs organization commonly known as the Skunk Works®, which is world renowned for creating breakthrough technologies and landmark aircraft. He is a Fellow of the American Institute of Aeronautics and Astronautics and of the Royal Aeronautical Society (RAeS). He earned a PhD in aerospace engineering from Georgia Institute of Technology in 1976 after earning a master’s degree in aeronautical engineering and a bachelor’s in electrical technology, both from the Indian Institute of Science in Bangalore, India.
Project support
This project was made possible in part through financial support from the University Libraries’ Open Education Initiative and the Kevin T. Crofton Department of Aerospace and Ocean Engineering.
Suggested citation
William H. Mason and Pradeep Raj, Lecture Notes on Configuration Aerodynamics (2026). CC BY-NC-SA 4.0. https://doi.org/10.21061/configurationaerodynamics.
View errata |
Report an error
Accessibility
Virginia Tech is committed to making its publications accessible in accordance with the Americans with Disabilities Act of 1990. The text, images, and links in the PDF versions of this text are tagged structurally and include alternative text, which allows for machine readability. Virginia Tech Publishing is continuously working to improve accessibility and welcomes any feedback from readers