Search NASA⌕ Search

SEARCH · Search NASA

Results for “decision trees”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
NASA NTRS2022

Albuquerque Urban Development: Enhancing Urban Cooling Interventions by Modeling Urban Forestry through NASA Earth Observations in Albuquerque, New Mexico

The City of Albuquerque, New Mexico is experiencing increasing urban heat island(UHI) effects, which impact the health, safety, and comfort of the community. In partnership with the City of Albuquerque Department of Environmental Health, Department of Parks and Recreation, and Let’s Plant Albuquerque!, this project used satellite Earth observations from April 2016–August 2022 to model increases in tree canopy within the City of Albuquerque to help combat the urban heat island in the city’s warmer areas over the next decade. Using Landsat 8’s Thermal Infrared Sensor (TIRS) and the Ecosystem Spaceborne Thermal Radiometer Experiment on the International Space Station (ECOSTRESS), along with the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) Urban Cooling and ENVI-Met models, the team modeled tree cover interventions and created land surface temperature maps. These outputs will help the city make data-driven decisions for their tree planting goal in a targeted approach.

Max Stewart↗
NASA NTRS2010

Exploiting the Free Landsat Archive for Operational Monitoring of Ecosystem Condition and Change Across the Chesapeake Bay Watershed

For the first time, all imagery acquired by the Landsat series of satellites is being made available by the USGS to users at no cost. This represents a key opportunity to use Landsat in a truly operational monitoring framework: large regions of the U.S. such as the Chesapeake Bay Watershed can now be analyzed using "wall-to-wall" imagery at timescales from approximately 1 month to several years. With the future launch of the Landsat Data Continuity Mission (LDCM) and Decadal Survey missions such as the hyperspectral HyspIRI, it is imperative to develop robust processing systems to perform annual ecosystem assessments over large regions such as the Chesapeake Bay. We have been working at NASA's Goddard Space Flight Center (GSFC) to develop an integrative framework for inserting 30m, annual, Landsat based data and derived products into the existing decision support system for the Bay, with a particular focus on ecosystem condition and changes over the entire watershed. The basic goal is to use a 'stack' of Landsat imagery with 40% or less cloud cover to produce multi-date (2005-2009 period), cloud/shadow/gap-free composited surface reflectance products that will support the creation of watershed scale land cover/ use products and the monitoring of ecosystem change across the Bay. Our scientific focus extends beyond the conventional definition of land cover (i.e. a classification of vegetation type) as we propose to monitor both changes in surface type (e.g. forest to urban), vegetation structure (e.g. forest disturbance due to logging or insect damage), as well as winter crop cover. These processes represent a continuum from large, interannual changes in land cover type, to subtler, intra-annual changes associated with short-term disturbance. The free Landsat data are being processed to surface reflectance and composited using the existing Landsat Ecosystem Disturbance Adaptive Processing System here at NASA/ GSFC, and land cover products (type, tree cover, impervious cover, winter cover) are being produced using well-established decision tree and regression tree algorithms. The goal of this session is to present the data products that we have been developing to the Bay science community and to discuss potential avenues for improvements and usage of the products for decision support.

BrowndeColstoun, Eric↗
NASA NTRS2001

Launch Vehicle Design Process: Characterization, Technical Integration, and Lessons Learned

Engineering design is a challenging activity for any product. Since launch vehicles are highly complex and interconnected and have extreme energy densities, their design represents a challenge of the highest order. The purpose of this document is to delineate and clarify the design process associated with the launch vehicle for space flight transportation. The goal is to define and characterize a baseline for the space transportation design process. This baseline can be used as a basis for improving effectiveness and efficiency of the design process. The baseline characterization is achieved via compartmentalization and technical integration of subsystems, design functions, and discipline functions. First, a global design process overview is provided in order to show responsibility, interactions, and connectivity of overall aspects of the design process. Then design essentials are delineated in order to emphasize necessary features of the design process that are sometimes overlooked. Finally the design process characterization is presented. This is accomplished by considering project technical framework, technical integration, process description (technical integration model, subsystem tree, design/discipline planes, decision gates, and tasks), and the design sequence. Also included in the document are a snapshot relating to process improvements, illustrations of the process, a survey of recommendations from experienced practitioners in aerospace, lessons learned, references, and a bibliography.

Blair, J. C.↗
DOE OSTI2025

LC-Opt: Benchmarking Reinforcement Learning and Agentic AI for End-to-End Liquid Cooling Optimization in Data Centers

Liquid cooling is critical for thermal management in high-density data centers with the rising AI workloads. However, machine learning-based controllers are essential to unlock greater energy efficiency and reliability, promoting sustainability. We present LC-Opt, a Sustainable Liquid Cooling (LC) benchmark environment, for reinforcement learning (RL) control strategies in energy-efficient liquid cooling of high-performance computing (HPC) systems. Built on the baseline of a high-fidelity digital twin of Oak Ridge National Lab's Frontier Supercomputer cooling system, LC-Opt provides detailed Modelica-based end-to-end models spanning site-level cooling towers to data center cabinets and server blade groups. RL agents optimize critical thermal controls like liquid supply temperature, flow rate, and granular valve actuation at the IT cabinet level, as well as cooling tower (CT) setpoints through a Gymnasium interface, with dynamic changes in workloads. This environment creates a multi-objective real-time optimization challenge balancing local thermal regulation and global energy efficiency, and also supports additional components like a heat recovery unit (HRU). We benchmark centralized and decentralized multi-agent RL approaches, demonstrate policy distillation into decision and regression trees for interpretable control, and explore LLM-based methods that explain control actions in natural language through an agentic mesh architecture designed to foster user trust and simplify system management. LC-Opt democratizes access to detailed, customizable liquid cooling models, enabling the ML community, operators, and vendors to develop sustainable data center liquid cooling control solutions.

Naug, Avisek [Hewlett Packard Enterprise]↗
NASA NTRS1980

AgRISTARS: Foreign commodity production forecasting. Corn/soybean decision logic development and testing

The development and testing of an analysis procedure which was developed to improve the consistency and objectively of crop identification using Landsat data is described. The procedure was developed to identify corn and soybean crops in the U.S. corn belt region. The procedure consists of a series of decision points arranged in a tree-like structure, the branches of which lead an analyst to crop labels. The specific decision logic is designed to maximize the objectively of the identification process and to promote the possibility of future automation. Significant results are summarized.

Dailey, C. L.↗
NASA NTRS1991

Learning classification trees

Algorithms for learning classification trees have had successes in artificial intelligence and statistics over many years. How a tree learning algorithm can be derived from Bayesian decision theory is outlined. This introduces Bayesian techniques for splitting, smoothing, and tree averaging. The splitting rule turns out to be similar to Quinlan's information gain splitting rule, while smoothing and averaging replace pruning. Comparative experiments with reimplementations of a minimum encoding approach, Quinlan's C4 and Breiman et al. Cart show the full Bayesian algorithm is consistently as good, or more accurate than these other approaches though at a computational price.

Buntine, Wray↗
NASA NTRS1990

An expert system for geologic mapping with imaging spectrometers

Techniques have been developed for the extraction and characterization of absorption features from visible and infrared reflectance spectra and an expert system has been designed, implemented, and successfully tested that allows automated identification of minerals based on their spectral characteristics. A suite of laboratory spectra of common minerals was analyzed and the absorption band characteristics tabulated and used to develop a generalized knowledge base for analysis of the reflectance spectra. A tree hierarchy was designed to emulate the decision process followed by an experienced analyst for analysis of laboratory and field reflectance spectra and aircraft imaging spectrometer spectra. Good results were obtained with the expert system for all three types of spectra, with the critical factor in the analysis being the signal-to-noise ratio of the spectral data.

Kruse, F. A.↗
NASA NTRS1994

Maximum-likelihood soft-decision decoding of block codes using the A* algorithm

The A* algorithm finds the path in a finite depth binary tree that optimizes a function. Here, it is applied to maximum-likelihood soft-decision decoding of block codes where the function optimized over the codewords is the likelihood function of the received sequence given each codeword. The algorithm considers codewords one bit at a time, making use of the most reliable received symbols first and pursuing only the partially expanded codewords that might be maximally likely. A version of the A* algorithm for maximum-likelihood decoding of block codes has been implemented for block codes up to 64 bits in length. The efficiency of this algorithm makes simulations of codes up to length 64 feasible. This article details the implementation currently in use, compares the decoding complexity with that of exhaustive search and Viterbi decoding algorithms, and presents performance curves obtained with this implementation of the A* algorithm for several codes.

Ekroot, L.↗
NASA NTRS2013

Ensemble Statistical Post-Processing of the National Air Quality Forecast Capability: Enhancing Ozone Forecasts in Baltimore, Maryland

An ensemble statistical post-processor (ESP) is developed for the National Air Quality Forecast Capability (NAQFC) to address the unique challenges of forecasting surface ozone in Baltimore, MD. Air quality and meteorological data were collected from the eight monitors that constitute the Baltimore forecast region. These data were used to build the ESP using a moving-block bootstrap, regression tree models, and extreme-value theory. The ESP was evaluated using a 10-fold cross-validation to avoid evaluation with the same data used in the development process. Results indicate that the ESP is conditionally biased, likely due to slight overfitting while training the regression tree models. When viewed from the perspective of a decision-maker, the ESP provides a wealth of additional information previously not available through the NAQFC alone. The user is provided the freedom to tailor the forecast to the decision at hand by using decision-specific probability thresholds that define a forecast for an ozone exceedance. Taking advantage of the ESP, the user not only receives an increase in value over the NAQFC, but also receives value for An ensemble statistical post-processor (ESP) is developed for the National Air Quality Forecast Capability (NAQFC) to address the unique challenges of forecasting surface ozone in Baltimore, MD. Air quality and meteorological data were collected from the eight monitors that constitute the Baltimore forecast region. These data were used to build the ESP using a moving-block bootstrap, regression tree models, and extreme-value theory. The ESP was evaluated using a 10-fold cross-validation to avoid evaluation with the same data used in the development process. Results indicate that the ESP is conditionally biased, likely due to slight overfitting while training the regression tree models. When viewed from the perspective of a decision-maker, the ESP provides a wealth of additional information previously not available through the NAQFC alone. The user is provided the freedom to tailor the forecast to the decision at hand by using decision-specific probability thresholds that define a forecast for an ozone exceedance. Taking advantage of the ESP, the user not only receives an increase in value over the NAQFC, but also receives value for

ozone↗
DOE OSTI2026

Feature Engineering and Ensemble Methods for Imbalanced ICS Intrusion Detection: Pipeline Audit and Constrained Evaluation

Industries are becoming increasingly connected and are more vulnerable to cyberattacks due to the widened attack surface. Industrial Control Systems (ICS) are among the most critical sectors that malicious actors can target, as such attacks can cause significant operational disruption and physical damage. It is imperative to detect such attacks as early as possible. This paper evaluates constraint-conditioned optimistic performance estimates for traditional ML models in ICS intrusion detection (i.e., estimates obtained under contiguous, non-shuffled temporal evaluation without test-set alteration, but with pre-split feature engineering that may introduce temporal leakage, due to dataset constraints). Our findings are threefold. First, we quantify how iterative feature engineering affects tree-based ensemble performance and examine how pipeline decisions (split strategy, sampling scope, and cleaning policy) can inflate or reduce reported IDS results under constraint-bound evaluation. Second, we compare intrinsic class-imbalance handling across ensemble models. Third, under our current pipeline constraints (including pre-split feature engineering), CatBoost achieves the best performance on Water Storage Tank (accuracy: 0.9831, class-1 F1: 0.9682), while Light- GBM achieves the best performance on Gas Pipeline (accuracy: 0.9618, class-1 F1: 0.9086).

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1967

Detection and estimation theory

Equalization of dispersive channels using decision feedback, and state variable estimation of sonar or seismic data in presence of pure delay

COMMUNICATION THEORY↗
NASA NTRS2003

A Dynamic Optimization Technique for Siting the NASA-Clark Atlanta Urban Rain Gauge Network (NCURN)

NASA satellites and ground instruments have indicated that cities like Atlanta, Georgia may create or alter rainfall. Scientists speculate that the urban heat island caused by man-made surfaces in cities impact the heat and wind patterns that form clouds and rainfall. However, more conclusive evidence is required to substantiate findings from satellites. NASA, along with scientists at Clark Atlanta University, are implementing a dense, urban rain gauge network in the metropolitan Atlanta area to support a satellite validation program called Studies of PRecipitation Anomalies from Widespread Urban Landuse (SPRAWL). SPRAWL will be conducted during the summer of 2003 to further identify and understand the impact of urban Atlanta on precipitation variability. The paper provides an. overview of SPRAWL, which represents one of the more comprehensive efforts in recent years to focus exclusively on urban-impacted rainfall. The paper also introduces a novel technique for deploying rain gauges for SPRAWL. The deployment of the dense Atlanta network is unique because it utilizes Geographic Information Systems (GIS) and Decision Support Systems (DSS) to optimize deployment of the rain gauges. These computer aided systems consider access to roads, drainage systems, tree cover, and other factors in guiding the deployment of the gauge network. GIS and DSS also provide decision-makers with additional resources and flexibility to make informed decisions while considering numerous factors. Also, the new Atlanta network and SPRAWL provide a unique opportunity to merge the high-resolution, urban rain gauge network with satellite-derived rainfall products to understand how cities are changing rainfall patterns, and possibly climate.

Shepherd, J. Marshall↗
DOE OSTI2024

A Behavior Tree Approach for Battery-Aware Inspection of Large Structures Using Drones

Electric multi-rotor drones have been used to inspect several structures, including large buildings and dams. In these inspections, energy consumption is a concern. To prevent the drone from running out of battery, commercial drones usually come back to their home position when the battery level reaches a minimum threshold. The pilots then need to replace the battery and use their own experience to restart the inspection mission approximately from where it ended before the drone returned home. Instead of relying on the human operator, in this paper, we automate this process using behavior trees, which is an effective way to perform autonomous mission control and supervision. By integrating battery management strategies into a behavior tree framework, this paper demonstrates the drone’s adaptive and resilient decision-making when confronted with limited power constraints. We implemented our methodology using a commercial drone and tested the proposed ideas in a photogrammetry-based inspection task.

42 ENGINEERING↗
NASA NTRSDate not supplied

Battery State-of-Health Aware Path Planning for a Mars Rover

A rover mission consists of visiting waypoints to gather scientific samples based on set requirements. However, rovers face operational uncertainties during the mission, affecting the performance of its electrical and mechanical components and overall mission success. Hence, it is critical to have a decision-making framework that is aware of the health state of the components when planning the path of the vehicle. In particular, battery degradation, and consequently the battery State of Health (SOH), can affect the optimality of decisions made by the autonomous system in the long term. This paper presents a decision-making system that incorporates information on the energy drawn from the battery (based on the vehicle’s velocity), terrain conditions, and model-based prognostic modules to assess the impact on the battery’s state of charge (SoC). The decision-making system was formulated as a Markov Decision Process (MDP) to reach the goal destination by sending commands in a determined amount of time while maintaining the battery SoC within the policy stated. The MDP problem was programmed using the open-source framework POMDPs.jl, which has a variety of online and offline solvers. To solve the MDP problem online, we used Monte Carlo Tree Search (MCTS). Results from simulations demonstrate the effect that battery degradation and charging plans have on decision-making.

Prognostics↗
NASA NTRS2014

Data Used in Quantified Reliability Models

Data is the crux to developing quantitative risk and reliability models, without the data there is no quantification. The means to find and identify reliability data or failure numbers to quantify fault tree models during conceptual and design phases is often the quagmire that precludes early decision makers consideration of potential risk drivers that will influence design. The analyst tasked with addressing a system or product reliability depends on the availability of data. But, where is does that data come from and what does it really apply to? Commercial industries, government agencies, and other international sources might have available data similar to what you are looking for. In general, internal and external technical reports and data based on similar and dissimilar equipment is often the first and only place checked. A common philosophy is "I have a number - that is good enough". But, is it? Have you ever considered the difference in reported data from various federal datasets and technical reports when compared to similar sources from national and/or international datasets? Just how well does your data compare? Understanding how the reported data was derived, and interpreting the information and details associated with the data is as important as the data itself.

DeMott, Diana↗
NASA NTRS2009

Probabilistic Risk Assessment for Decision Making During Spacecraft Operations

Decisions made during the operational phase of a space mission often have significant and immediate consequences. Without the explicit consideration of the risks involved and their representation in a solid model, it is very likely that these risks are not considered systematically in trade studies. Wrong decisions during the operational phase of a space mission can lead to immediate system failure whereas correct decisions can help recover the system even from faulty conditions. A problem of special interest is the determination of the system fault protection strategies upon the occurrence of faults within the system. Decisions regarding the fault protection strategy also heavily rely on a correct understanding of the state of the system and an integrated risk model that represents the various possible scenarios and their respective likelihoods. Probabilistic Risk Assessment (PRA) modeling is applicable to the full lifecycle of a space mission project, from concept development to preliminary design, detailed design, development and operations. The benefits and utilities of the model, however, depend on the phase of the mission for which it is used. This is because of the difference in the key strategic decisions that support each mission phase. The focus of this paper is on describing the particular methods used for PRA modeling during the operational phase of a spacecraft by gleaning insight from recently conducted case studies on two operational Mars orbiters. During operations, the key decisions relate to the commands sent to the spacecraft for any kind of diagnostics, anomaly resolution, trajectory changes, or planning. Often, faults and failures occur in the parts of the spacecraft but are contained or mitigated before they can cause serious damage. The failure behavior of the system during operations provides valuable data for updating and adjusting the related PRA models that are built primarily based on historical failure data. The PRA models, in turn, provide insight into the effect of various faults or failures on the risk and failure drivers of the system and the likelihood of possible end case scenarios, thereby facilitating the decision making process during operations. This paper describes the process of adjusting PRA models based on observed spacecraft data, on one hand, and utilizing the models for insight into the future system behavior on the other hand. While PRA models are typically used as a decision aid during the design phase of a space mission, we advocate adjusting them based on the observed behavior of the spacecraft and utilizing them for decision support during the operations phase.

dynamic fault trees↗
NASA NTRSDate not supplied

Utilizing NASA Earth Observations to Evaluate Urban Tree Canopy and Land Surface Temperature for Green Infrastructure Development and Urban Heat Mitigation in Huntsville, AL

Huntsville, Alabama’s population has grown by 10.8% since 2010, due in part to the city’s advancing engineering industry. Rapid urban growth negatively impacts the environment by decreasing tree canopy cover and increasing impervious surface cover, which can intensify the urban heat island effect. To examine the impacts of this urban growth on the environment, the team partnered with the City of Huntsville to utilize Landsat 5 Thematic Mapper (TM), Landsat 8 Operational Land Imager (OLI), Landsat 8 Thermal Infrared Sensor (TIRS), Terra Moderate Resolution Imaging Spectroradiometer (MODIS) and the International Space Station’s Global Ecosystems Dynamic Investigation (GEDI) and ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS). The team utilized these Earth observations in combination with ancillary datasets to create a suite of end products to assist in mitigating the effects of extreme heat due to urban expansion and tree canopy loss. Rasters of annual land surface temperature (LST) was calculated in Google Earth Engine from 2010 to 2019. The team derived land cover classes through supervised and threshold classification methods to distinguish trees, other vegetation types, impervious surfaces, and water. From 2010 to 2019, LST increased approximately 4 °F for all census tracts within the city and the total amount of tree cover increases less than 3%. The findings will aid the city in future decision-making processes by indicating areas that would benefit from increased green infrastructure.

Greta Paris↗