100039 research outputs found
Sort by
Advancing Multi-Agent Robotics Simulations Through Heterogeneous Reinforcement Learning in IsaacLab
Get PDFRobots increasingly operate in collaborative teams across domains such as search-and- rescue, warehouse automation, and autonomous driving—scenarios that demand advanced coordination strategies enabled by multi-agent reinforcement learning (MARL). However, existing simulation frameworks often struggle to balance realism, speed, and scalability, especially when supporting diverse, heterogeneous robot teams. This research extends Isaac Lab, a high-performance robotics simulator, by integrating heterogeneous-agent reinforcement learning (HARL) capabilities. The result is a flexible and GPU-accelerated platform for training both homogeneous and heterogeneous robot teams in complex, physics-based environments. These enhancements significantly narrow the gap between simulation and real-world deployment for multi-robot systems
Component Analysis of Interactive Computer Training to Teach Paraprofessionals to Deliver Behavioral Skills Training
Get PDFIn education and Applied Behavior Analysis, there is growing demand for well-trained entry-level individuals to deliver behavior interventions. Interactive Computer Training allows trainees to access online instruction with video demonstrations, interactive tasks, and feedback, without a live instructor. This training approach could save valuable training time and money. Prior researchers demonstrated these types of training can teach trainees to correctly use behavior interventions. There are gaps in our understanding of Interactive Computer Trainings that limit its successful use. Few studies measured cost-effectiveness, identified the effective features of Interactive Computer Training (e.g., videos, quizzes, feedback), or trained users to deliver effective training practices. The purpose of this study was to reproduce and extend prior research by assessing the active features of Interactive Computer Training to teach training skills with four undergraduate paraprofessionals. This was done by measuring the reliability of training skill use, generalization to an untrained task, long-term skill maintenance, and training development requirements. The results were similar to prior research showing Interactive Computer Training can improve training skill delivery. The features necessary for correct target skill delivery varied across participants, but half required the full training package and performance feedback. Computer training programs that include detailed video examples and knowledge-quizzes, appear to be the most effective for trainee success. Time and cost could be a restrictive factor for training settings that can hinder computer training use and should be a consideration in practice and study. Limitations in the study design and measurement limit the suggestions that can be made regarding the effective components of computer training and its effectiveness to teach training skills. More work must be done to understand how these features and their design can impact skill learning, so more efficient training packages can be designed. Studying these training packages within large-scale settings that require plentiful training efforts (e.g., school districts, organization-wide) could show that Interactive Computer Training can train large groups of people and beneficial skills to directly improve learner success
Deterring Gray Zone Activity in a Strategic Flashpoint: U.S. Freedom of Navigation Operations and China’s Maritime Militias
Get PDFIn today’s complex world, nations often use irregular forces, such as militias, to achieve political goals without escalating conflicts. This study examines how the United States of America uses its Navy’s Freedom of Navigation Operations in the South China Sea to influence the behavior of Chinese maritime militias. These militias are involved in activities such as harassment, assault, and other aggressive or deceptive actions. By analyzing data from 2012 to 2024, this research discovered 239 incidents and discovers that these militias reduce their confrontational activities during Freedom of Navigation Operations to avoid direct conflict while continuing their broader objectives in the region.
This research is critical because the South China Sea is a hotspot for international tension and competition. Understanding how maritime militias operate and respond to U.S. naval actions offers valuable insights for policymakers. The findings can help governments develop strategies to reduce risks, strengthen their defenses, and prevent future conflicts. By shedding light on this issue, the study contributes to global efforts to maintain peace and stability in one of the world’s most contested regions
Exploring Variations in High-Resolution Downstream Hydraulic Geometry of the Logan River
Get PDFRivers are dynamic systems that change shape due to natural events like floods, landslides, and wildfires, as well as human-made structures like bridges. These changes create unique and complex river patterns. However, measuring these changes accurately is challenging because traditional field surveys are time-consuming and often focus on easily accessible areas, which may not represent the whole river. Thanks to new high-resolution mapping technology and detailed river flow records, we can now see these changes in greater detail than ever before.
In this study, we focused on the Logan River, which flows through a canyon with a mix of rocky and sediment-filled areas. Using advanced mapping tools, we measured the river\u27s width along a 45 km stretch and checked these measurements against field surveys to ensure accuracy. This approach allows us to identify sections of the river that show consistent patterns and those that deviate due to factors like incoming streams or human impacts like roads. Understanding these variations helps us to better plan for infrastructure projects, river restoration, and managing natural hazards. Our method provides a new way to understand how changes in the environment affect rivers, which can lead to smarter, more resilient design and planning
Evaluating Elk Distribution and Conflict Under Proposed Management Alternatives at the National Elk Refuge in Jackson, Wyoming
Get PDFWe evaluated measurable attributes describing the current and future distribution of Cervus elaphus canadensis (elk) across a region surrounding Jackson, Wyoming, for five feedground management alternatives proposed by the U.S. Fish and Wildlife Service as a revision to 2007 Bison and Elk Management Plan of the National Elk Refuge. A resource selection function evaluated measurable attributes of interest to managers, including elk use of private property and sensitive habitat types at monthly timesteps and varying winter conditions. The study area boundaries were created through an expert elicitation process and consist of the Jackson Elk Herd Unit, Grand Teton National Park, the National Elk Refuge, and the northern third of the Fall Creek Elk Herd Unit. For each of the five alternatives, we distributed monthly elk numbers calculated in a concurrent analysis that simulated chronic wasting disease dynamics in this system for 20 years. Measurable attributes representing potential elk use of (1) private property, (2) cattle properties as an index of Brucella abortus risk, and sensitive habitats consisting of (3) Populus tremuloides Michx. (quaking aspen), (4) Populus angustifolia E. James (narrowleaf cottonwood), and (5) Salix L. (willow) in core winter use all areas all closely followed the declines of elk abundance projected by the elk chronic wasting disease model. After 20 years, the continue feeding alternative ranked most favorably in terms of limiting elk days on private property and reducing brucellosis risk from elk to cattle because this alternative concentrated elk on the National Elk Refuge and resulted in the lowest elk population sizes. However, other management alternatives, including increase harvest and reduce feeding, tended to limit elk use of sensitive quaking aspen, narrowleaf cottonwood, and willow habitats during winter (December–April)
Timber Harvesting was the Most Important Factor Driving Changes in Vegetation Composition, As Compared to Climate and Fire Regime Shifts, In the Mixedwood Temperate Forests of Temiscamingue Since AD 1830
Get PDFContext The vegetation composition of northeastern North American forests has significantly changed since pre-settlement times, with a marked reduction in conifer-dominated stands, taxonomic and functional diversity. These changes have been attributed to fire regime shifts, logging, and climate change.
Methods In this study, we disentangled the individual effects of these drivers on the forest composition in southwestern Quebec from 1830 to 2000 by conducting retrospective modelling using the LANDIS-II forest landscape model. The model was run based on pre-settlement forest composition and fire history reconstructions, historical timber harvest records, and climate reanalysis data. We compared counterfactual scenarios excluding individual factors to a baseline historical scenario.
Results and Conclusions Our results indicated that timber harvesting had the greatest impact on forest dynamics over the past centuries. In the absence of timber harvesting, pre-settlement species abundances were largely maintained, preserving key functional traits like fire and shade tolerance that contribute to ecosystem resilience. Increased fire activity during the settlement period contributed to the increase of early-successional aspen (Populus tremuloides), but timber harvesting played the dominant role. Fire exclusion had no influence on vegetation composition, suggesting mesophication unfolds over longer timescales than those captured in this study. Climate change, characterized by modest increases in temperature and precipitation, had a minor effect on vegetation shifts, as increased precipitation might have mitigated the adverse effects of rising temperatures. However, future climate change is projected to become a more significant driver of forest composition. These findings underscore the importance of forest restoration and continued research on past forest dynamics to better understand current and future changes
Ultra-Low Power ADCs for Charge Detection Mass Spectrometry of Martian Dust
No full textThis paper presents a complete system integration strategy for a proven Charge Detection Mass Spectrometry (CDMS) instrument designed to analyze charged dust particles on Mars. To enable deployment in a power-constrained rover environment, an ultra-low supply voltage (ULV) analog-to-digital converter (ADC) is required. ADC architectures operating below 0.21 V are compared to identify candidates that meet the system\u27s stringent power and performance demands. A voltage-controlled oscillator (VCO)-based ADC is proposed and for ULV operation. Circuit-level implementations of VCO topologies are designed and evaluated in the TSMC 28 nm CMOS process using Cadence simulations. Based on simulation results, a VCO-CDAC topology is selected for future integration into the full ADC architecture
Reconstruction of CAD Models Into Analysis-Ready Single Trimmed Surfaces With Aerospace Applications
No full textClassical methods by which computer-aided design (CAD) geometries are represented for both design and analysis involve meshing. However, the current process of converting a CAD model into an analysis-suitable surface mesh takes a significant amount of time and labor [23, 4, 18], results in a faceted representation of the original smooth geometry, and generates representations that are prohibitively expensive for use in high-order explicit dynamics computations [35, 9].
This research focuses on accurately and efficiently rebuilding given CAD surfaces or meshes, such as shell-like components of aeronautical structures or automotive vehicles, into smooth, single trimmed spline surface approximations (rather than typical piecewise-linear approximations), making them suitable for use in isogeometric analysis (IGA). Spline-based discretizations exhibit higher accuracy per degree of freedom than their piecewise-faceted counterparts [40, 16, 8, 6], possess high-order continuity (e.g., for use in Kirchhoff-Love shell analysis), and have potential for much larger stable time steps in explicit dynamics analyses than traditional finite-element methods [9, 20], which make them a compelling alternative to traditional piecewise-linear alternatives.
Our proposed method accurately represents and automatically reconstructs these structures by converting a CAD geometry into a feature-aware triangulation, which is then flattened using theory from conformal or differential geometry to define the parametric domain and trimming curves of the intended spline. The flattened geometry is then extended to fill its bounded parametric domain, with its extended vertices given a spatial position by minimizing their biharmonic energy. Subsequently, the bijection between this flattened geometry and its original spatial representation is then used to inform a mapping of a basis spline (B-spline) back into the spatial domain. Once the spatial surface is achieved, it is trimmed using the boundary of the original geometry, thereby reconstructing the intended geometry as a single trimmed B-spline surface. Resulting spline geometries are ultimately used for structural engineering analysis to demonstrate their sustainability for analysis
Analysis Ready Aerospace Wing Assembly Reconstruction Using Parameterization
No full textAccurate analysis of structures in the design process are often labor intensive and computationally expensive. Isogeometric analysis (IGA) provides potential to better connect the design through analysis process and converge to accurate results more quickly. Here, methods are proposed to help in the goal of automating the process of converting design geometry to an analysis suitable model for IGA. This is done by using tools previously built in combination with a motorcycle graph algorithm to extract quadrilateral layouts of the intended geometry. This method will be demonstrated by reconstructing and analyzing an airplane structure
Data Reduction and Photometry for Surface Brightness Fluctuation Measurements
No full textAmong the most pressing issues in modern cosmology is the Hubble tension, in which early-universe measurements of H0 differ from late-universe measurements by \u3e 5σ. This tension, now verging on crisis, implies continued systematic errors in either early or late universe measurement, or deeper issues with the widely accepted ΛCDM model of cosmology. To reduce systematic uncertainty, we have been developing an independent distance ladder relying on surface brightness fluctuations (SBF) and tip of the red-giant-branch (TRGB) star calibrators with data from the Hubble Space Telescope (HST) and James Webb Space Telescope (JWST). Before distance measurements can be made, data reduction and photometry must be performed