USMA Digital Commons (United States Military Academy, West Point)
Not a member yet
1355 research outputs found
Sort by
Robust Decision-Making in the Internet of Battlefield Things Using Bayesian Neural Networks
No full textThe Internet of Battlefield Things (IoBT) is a dynamically composed network of intelligent sensors and actuators that operate as a command and control, communications, computers, and intelligence complex-system with the aim to enable multi-domain operations. The use of artificial intelligence can help transform the IoBT data into actionable insight to create information and decision advantage on the battlefield. In this work, we focus on how accounting for uncertainty in IoBT systems can result in more robust and safer systems. Human trust in these systems requires the ability to understand and interpret how machines make decisions. Most real-world applications currently use deterministic machine learning techniques that cannot incorporate uncertainty. In this work, we focus on the machine learning task of classifying vehicles from their audio recordings, comparing deterministic convolutional neural networks (CNNs) with Bayesian CNNs to show that correctly estimating the uncertainty can help lead to robust decision-making in IoBT
Test-Retest Reliability of Concussion Baseline Assessments in United States Service Academy Cadets: A Report from the National Collegiate Athletic Association (NCAA)-Department of Defense (DoD) CARE Consortium.
No full textOBJECTIVE: In response to advancing clinical practice guidelines regarding concussion management, service members, like athletes, complete a baseline assessment prior to participating in high-risk activities. While several studies have established test stability in athletes, no investigation to date has examined the stability of baseline assessment scores in military cadets. The objective of this study was to assess the test-retest reliability of a baseline concussion test battery in cadets at U.S. Service Academies.
METHODS: All cadets participating in the Concussion Assessment, Research, and Education (CARE) Consortium investigation completed a standard baseline battery that included memory, balance, symptom, and neurocognitive assessments. Annual baseline testing was completed during the first 3 years of the study. A two-way mixed-model analysis of variance (intraclass correlation coefficent (ICC)3,1) and Kappa statistics were used to assess the stability of the metrics at 1-year and 2-year time intervals.
RESULTS: ICC values for the 1-year test interval ranged from 0.28 to 0.67 and from 0.15 to 0.57 for the 2-year interval. Kappa values ranged from 0.16 to 0.21 for the 1-year interval and from 0.29 to 0.31 for the 2-year test interval. Across all measures, the observed effects were small, ranging from 0.01 to 0.44.
CONCLUSIONS: This investigation noted less than optimal reliability for the most common concussion baseline assessments. While none of the assessments met or exceeded the accepted clinical threshold, the effect sizes were relatively small suggesting an overlap in performance from year-to-year. As such, baseline assessments beyond the initial evaluation in cadets are not essential but could aid concussion diagnosis
Thinking Like a Futurist: Investigating the Theories and Processes of Threatcasting Post-Analysis
No full textThreatcasting is a foresight methodology that examines the worst of potential future changes by imagining and crafting a fictional (but very plausible) story of a person, in a detailed setting, experiencing a threat. In this dissertation, I investigate the processes and techniques of threatcasting, focused primarily on the post-analysis phase, and demonstrate it as an open methodology that can embrace varied ways to analyze raw data and seek conclusions. I incorporate best practices of narrative and thematic analysis, qualitative analysis, grounded theory, and hypothesis-driven theories of inquiry. I use interviews from futurists trained on threatcasting ways of thinking and compare two case studies - one using a grounded theory approach on the future of weapons of mass destruction and cyberspace and the other using a hypothesis-driven approach on the future of extremism - to investigate the efficacy of different theoretical approaches to analysis. I introduce definitions of novelty and ways to assess how a novel finding may have more impact on the future than it appears at first glance. Often, this impact comes more from what is not present in threat scenarios than what is included. Finally, I illustrate how threatcasting, as a practice, is a valuable contribution to those in a position to be responsible architects of a better future
Comparing Feedback Linearization and Adaptive Backstepping Control for Airborne Orientation of Agile Ground Robots using Wheel Reaction Torque
No full textIn this paper, two nonlinear methods for stabilizing the orientation of a Four-Wheel Independent Drive and Steering (4WIDS) robot while in the air are analyzed, implemented in simulation, and compared. AGRO (the Agile Ground Robot) is a 4WIDS inspection robot that can be deployed into unsafe environments by being thrown, and can use the reaction torque from its four wheels to command its orientation while in the air. The goal of this work is to decrease the stabilization time and reject disturbances using nonlinear control methods. Model-based Feedback Linearization (FL) was added to PD control to compensate for nonlinear dynamics. However, with external disturbances, model uncertainty, and sensor noise the FL+PD controller does not guarantee stability. As an alternative, a backstepping controller was designed based on Lyapunov analysis with adaptive compensation for external disturbances, model uncertainty, and sensor offset. A simulation was written using the full nonlinear dynamics of AGRO in an isotropic steering configuration in which control authority over its pitch and roll are equalized. The PD+FL control method was compared to the backstepping control method using the same initial conditions in simulation. Both the backstepping controller and the PD+FL controller stabilized the system within 250 ms. The adaptive backstepping controller was also able to compensate for offset noisy sinusoidal disturbances through an adaptation law
Gender, Body Mass and Lean Body Mass Relationships on A Robust Fitness Challenge
No full textRobust fitness testing {Cadet Fitness Challenge (CFC)} is an important entity for physical assessment & future military operations. PURPOSE: Investigate relationships of gender, body mass, & lean body mass on performance during the CFC at a U.S. Service Academy
Toward Safe Decision-Making via Uncertainty Quantification in Machine Learning
No full textThe automation of safety-critical systems is becoming increasingly prevalent as machine learning approaches become more sophisticated and capable. However, approaches that are safe to use in critical systems must account for uncertainty. Most real-world applications currently use deterministic machine learning techniques that cannot incorporate uncertainty. In order to place systems in critical infrastructure, we must be able to understand and interpret how machines make decisions. This need is so that they can provide support for human decision-making, as well as the potential to operate autonomously. As such, we highlight the importance of incorporating uncertainty into the decision-making process and present the advantages of Bayesian decision theory. We showcase an example of classifying vehicles from their acoustic recordings, where certain classes have significantly higher threat levels. We show how carefully adopting the Bayesian paradigm not only leads to safer decisions, but also provides a clear distinction between the roles of the machine learning expert and the domain expert
In Situ Exploration of Soil Lead in Residential Communities Using X-Ray Fluorescence and Geospatial Visualization
No full textLead contamination in soil is a human health hazard common in residential communities that pre-date regulatory bans on lead in both gasoline and paints. New remote sensing tools allow for quicker and more affordable sampling, but there is still a challenge in interpreting the data, visualizing the results, and communicating the relevance for response and remediation. Our work builds on previous studies analyzing soil lead concentrations at West Point, NY. The federal installation and college campus hosts residential neighborhoods with historic homes that were painted with lead paint in the past and are adjacent to high traffic roadways. Previous research established several areas where the lead concentrations significantly exceeded the U.S. Environmental Protection Agency (EPA) recommended safe concentrations for soil, but further exploration was necessary to refine those results. We targeted one location where a 2019 measurement indicated lead in excess of 1200 mg/kg. We used an X-Ray Fluorescence (XRF) meter to collect 73 soil lead concentrations between the road and the home, logging the locations using ArcGIS Collector. Results indicate localized lead concentrations with distinctive patterns that may provide clues to the origin of the contamination. Our analysis suggests that in situ measurements are effective to characterize concentrations but conclusions on the severity of lead contamination should not be made using widely spaced transect investigations. The XRF, combined with geospatial visualization methods, is a quick and inexpensive way to investigate neighborhood-scale soil lead contamination and refine the potential remediation response
A Celebration of West Point Authors, January - June 2021
Get PDFHighlighting the 318 collected works of scholarship published and presented between January - June 2021.https://digitalcommons.usmalibrary.org/books/1041/thumbnail.jp
One Team, One Mission: Faculty and Department Development
No full textThis presentation argues that effective faculty development occurs when faculty are embedded in a healthy, team-oriented department, where individuals feel safe and welcome. There is a body of evidence that the primary role of a leader is the development and maintenance of these team-oriented conditions. Dan Coyle in The Culture Code makes the argument that good leaders prioritize building a sense of belonging on their teams, and they do so by structuring and leading the department in ways that facilitate unit social identity, and which is congruent with developing a sense of self-identity that is consistent with best practices of faculty life. When successful the leader is then in a strong position to help the department establish a sense of purpose and mission that facilitates proficiency and creativity on the part of each individual in the group. Leadership in this environment is seen as an interactive process rather than a thing to be had, and leaders who lead well are seen as fulfilling certain roles for the group: as an in-group prototype of group values and behaviors, as an in-group culture champion, as an entrepreneur of the group\u27s social identity, and is an embedder of identity to new members of the group (Haslam, et al, 2011). This presentation will provide the theoretical background for these leadership processes, and provide examples of kinesiology leaders who have been, and are, successful in creating environments where faculty development is most likely to occur