11115 research outputs found
Sort by
Comparing YOLOv8 to YOLOv5 for Pose Estimation Supporting Automated Aerial Refueling
Get PDFThis work used the newly released YOLO version 8 Object Detection as a feature detector for a monocular pose estimation pipeline and showed up to an 83.6% reduction in translation error when compared to YOLOv5. YOLOv8’s new Pose task’s utility as a feature detector was also explored. It performed consistently worse than the Object Detection task, with typical translation error magnitudes between 3 and 8 times worse than Object Detection. Finally, we investigated the use of constant-sized bounding box labels for Object Detection. Previous approaches have used Bounding Box Corrections (BBC). We found that using constant sized labels increases the number of accurate predictions made by a pose prediction pipeline by up to 11%, while also reducing label generation complexity
An Analysis of China’s Perceived Geographic Locations of Interest by Use of Value Informed Facility Location Models
Get PDFThis research examines China and derives insights specific to it and the First Island Chain and the Second Island Chain. In doing so, this research demonstrates a methodology to examine other competitors and their geostrategic interests. In the first phase of analysis, it develops a value hierarchy to depict objectives within subregions of the area of interest and considers four alternative weightings of the value hierarchy. In the second phase of analysis, it applies four location-covering models to assess how the competitor would emplace a range of limited resources to deter and/or control points of interest. Results indicate that land-based capabilities are effective and that sea-based facilities are necessary for coverage
Comparative Analysis of Satellite Battery Technologies: A Cross-Sectional and Temporal Study Within and Across Battery Types
Get PDFTrends in costs for batteries incorporated in satellite electrical power systems (EPS) have been identified by Space Systems Command (SSC) within the Unmanned Space Vehicle Cost Model (USCM) database. A subset of pre-1980s data consisting of Nickel-Cadmium battery costs was excluded from battery cost estimating relationship (CER) development as a result of this determination. This research aims to ascertain the impacts of these trends on satellite cost estimating practices by assessing developments in battery costs over time, comparing normalized Nickel-Cadmium (NiCd) / NickelHydrogen (NiH2) / Lithium-Ion (Li-Ion) battery costs, identifying differences between the NiH2 and Li-Ion modern battery types, and proposing a strategy for handling battery cost data in CER development. Statistical tests employed to evaluate the exclusion of NiCd data produced strong evidence supporting this decision. Comparison tests between NiH2 and Li-Ion datasets demonstrated statistically similar trends providing justification to combine the two datasets for modeling purposes. The analysis of battery costs over time showed inconclusive results due to data limitations. Lastly, the tested hypothesis of a negative correlation between battery costs and technology maturity was not supported, challenging prevailing assumptions
A Statistical Analysis of Phased Construction Projects
Get PDFThe construction industry is known for exceeding target budgets and schedules, and Department of Defense projects are no exception. The Air Force Installation & Mission Support Center (AFIMSC) manages a budget of $3.14 billion to repair, alter, demolish, and maintain Air Force facilities. However, an ever-increasing backlog of maintenance and repair work and a consistently underfunded repair program pose an increased financial risk to the sustainability of the construction and repair program. Phasing large construction efforts is a potential strategy that has recently been used to yield efficiencies by AFIMSC program managers. This study presents a statistical analysis of phased construction projects within AFIMSC\u27s centralized budget to determine if project phasing is a viable solution to help curb the cost overruns and help improve overall portfolio performance. The study uses descriptive statistics to analyze the cost and funding of projects and discusses the potential benefits and drawbacks of phased construction. The study concludes that project phasing can be a viable solution to improve project efficiency while keeping cost overruns in check
Characterization of Per- and Polyflouroalkyl Substances in Aqueous Filmforming Foam Wash Waters
Get PDFThe fiscal year 2020 National Defense Authorization Act mandated the publication of a military specification for a fluorine-free firefighting foam available for use no later than October 1, 2023. Replacing AFFF with a fluorine-free alternative creates the issue of decontaminating current firefighting infrastructure of residual PFAS without replacing these systems. To eliminate outstanding PFAS, the proposed solution is to drain the AFFF from current fire suppression systems and flush the plumbing three times with uncontaminated water. This study tested the efficacy of the triple rinse method by conducting scaled down experiments on a variety of plumbing components from a retired aircraft hangar fire suppression system. The components tested included straight pipe sections, elbows, and proportioners. Pre- and post-rinse samples were analyzed using a scanning electron microscope (SEM) coupled with energy dispersive xray spectroscopy (EDS) and a model was created to analyze thermodynamic favorability among PFAS and metal oxides. Results indicate that the triple rinse method is capable of partially removing certain PFAS and that metal oxides may contribute to desorption kinetics. To this author’s awareness, this is the first research to study PFAS interactions with plumbing surfaces through the combination of experimental observations and density functional theory. The sum of these results is expected to inform decontamination protocols for firefighting infrastructure containing AFFF
Predictive Analytics for Military Construction Overruns: A Machine Learning Approach
Get PDFThis research focused on analyzing cost and schedule overruns in a specific set of Air Force MILCON projects. Using data from the Built Infrastructure Common Operating Picture (BICOP), several advanced machine learning techniques, including LASSO regression and Random Forest, were applied to uncover patterns and build predictive models. The work sought to enhance the decision-making capabilities and efficiency of project managers within the Air Force through data-driven insights. A key finding from the models indicated that location was the most important factor in influencing a project’s tendency toward cost and schedule overrun. The best model for predicting schedule overrun achieved an accuracy of 82%, which was 20% higher than the no-information rate. In contrast, the best model in predicting cost overrun attained an accuracy of 70%, which was 11% higher than the no-information rate. Although the predictive models demonstrated some level of predictive capability and exceeded the no-information rate, they were not consistently accurate enough to be recommended as an aid for decision-making. The findings suggest that while machine learning can provide valuable insights into factors influencing project overruns, further refinement and testing are needed before these models can be considered reliable for use to flag early signs of potential construction project delays or cost increases
Integrating EOD and Readiness Competencies into the CE CGO Competency Framework
Get PDFIn the last few years, the strategic focus of the United States defense posturing has shifted from legacy conflict in the Middle East to focusing on the Pacific with increased emphasis that Airmen must be educated and trained to deter or prevail in any conflict. To achieve this goal, the Department of Air Force (DAF) has evolved to utilizing a competency-based education framework with the Civil Engineer career field adopting this strategy and releasing the first version of their occupational competencies in 2020. However, DAF force posturing has evolved since 2020. With the new deployment system, Agile Combat Employment scheme of maneuver and the current threat posture, Explosive Ordnance Disposal (EOD and Readiness and Emergency Management (R&EM) specialists will be critical for guiding instillations though disruptive events to generate combat air power. The primary objective of this research was to determine how relevant the current competencies are to EOD and R&EM capabilities plus how the current competencies could be improved to better capture these efforts. Two expert panels were utilized to determine applicability of the current competencies in the Career Field and Education Training Plan (CFETP), to generate new concepts through a thematic analysis and to achieve consensus on the new concepts to be recommended for addition to the Civil Engineer Company Grade Officer Competency Framework. The analysis identified 44 concepts to be recommended for inclusion in the next CFETP update
Sensor-Based Vehicle Classification Using Machine Learning
Get PDFThis research investigates the classification of vehicles into heavy and light categories using acoustic, seismic, and magnetic sensor data. The effectiveness of using frequency domain data and classical machine learning techniques, is compared with the effectiveness of using time-series data and neural networks. The primary aim in doing so was to understand if modern neural network architectures could effectively remove the need for more traditional frequency based signals processing. A significant deliverable of this thesis was the feature importance determined for each of the three phenomenological types found within the data (acoustic, seismic, and magnetic). By analyzing the importance of features derived from acoustic, seismic, and magnetic sensors, the research provided insights into which sensor types and their specific characteristics were most critical for distinguishing between heavy and light vehicles
Sparse Sensor Placement Optimization for Prediction of Angle of Attack with Artificial Hair-Cell Airflow Microsensors
No full textArrays of bioinspired artificial hair-cell airflow velocity sensors can enable flight-by-feel of small, unmanned aircraft. Natural fliers - bats, insects, and birds - have hundred or thousands of velocity sensors distributed across their wings. Aircraft designers do not have this luxury due to size, weight, and power constraints. The challenge is to identify the best locations for a small set of sensors to extract relevant information from the flow field for the prediction of flight control parameters. In this paper, we introduce the data-reducing Sparse Sensor Placement Optimization for Prediction algorithm which locates near-optimal sensor placement on airfoils and wings. For two or more sensors this algorithm finds a set of sensor locations (design point) which predicts angle of attack to within 0.10 degrees and ranks within the top 1 percent of all possible design points found by brute force search. We demonstrate this algorithm on several variations of airfoil sections of infinite and finite wings in clean and noisy data, evaluate model sensitivities, and show that the algorithm can be used to identify an appropriate number of sensors for a given accuracy requirement. Applications for this algorithm are explored for aircraft design and flight-by-feel control