Search NASA⌕ Search

SEARCH · Search NASA

Results for “learning efficiency”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRS2019

Using Machine Learning to Predict Core Sizes of High-Efficiency Turbofan Engines

With the rise in big data and analytics, machine learning is transforming many industries. It is being increasingly employed to solve a wide range of complex problems, producing autonomous systems that support human decision-making. For the aircraft engine industry, machine learning of historical and existing engine data could provide insights that help drive for better engine design. This work explored the application of machine learning to engine preliminary design. Engine core-size prediction was chosen for the first study because of its relative simplicity in terms of number of input variables required (only three). Specifically, machine-learning predictive tools were developed for turbofan engine core-size prediction, using publicly available data of two hundred manufactured engines and engines that were studied previously in NASA aeronautics projects. The prediction results of these models show that, by bringing together big data, robust machine-learning algorithms and automation, a machine learning-based predictive model can be an effective tool for turbofan engine core-size prediction. The promising results of this first study paves the way for further exploration of the use of machine learning for aircraft engine preliminary design.

Tong, Michael T.↗
NASA NTRS2019

Using Machine Learning to Predict Core Sizes of High-Efficiency Turbofan Engines

With the rise in big data and analytics, machine learning is transforming many industries. It is being increasingly employed to solve a wide range of complex problems, producing autonomous systems that support human decision-making. For the aircraft engine industry, machine learning of historical and existing engine data could provide insights that help drive for better engine design. This work explored the application of machine learning to engine preliminary design. Engine core-size prediction was chosen for the first study because of its relative simplicity in terms of number of input variables required (only three). Specifically, machine-learning predictive tools were developed for turbofan engine core-size prediction, using publicly available data of two hundred manufactured engines and engines that were studied previously in NASA aeronautics projects. The prediction results of these models show that, by bringing together big data, robust machine-learning algorithms and data science, a machine learning-based predictive model can be an effective tool for turbofan engine core-size prediction. The promising results of this first study paves the way for further exploration of the use of machine learning for aircraft engine preliminary design.

Core Size↗
NASA NTRS2018

Quantum Machine Learning

Quantum computing promises an unprecedented ability to solve intractable problems by harnessing quantum mechanical effects such as tunneling, superposition, and entanglement. The Quantum Artificial Intelligence Laboratory (QuAIL) at NASA Ames Research Center is the space agency's primary facility for conducting research and development in quantum information sciences. QuAIL conducts fundamental research in quantum physics but also explores how best to exploit and apply this disruptive technology to enable NASA missions in aeronautics, Earth and space sciences, and space exploration. At the same time, machine learning has become a major focus in computer science and captured the imagination of the public as a panacea to myriad big data problems. In this talk, we will discuss how classical machine learning can take advantage of quantum computing to significantly improve its effectiveness. Although we illustrate this concept on a quantum annealer, other quantum platforms could be used as well. If explored fully and implemented efficiently, quantum machine learning could greatly accelerate a wide range of tasks leading to new technologies and discoveries that will significantly change the way we solve real-world problems.

Biswas, Rupak↗
NASA NTRS2016

A Quantum-Assisted Algorithm for Sampling Applications in Machine Learning

An increase in the efficiency of sampling from Boltzmann distributions would have a significant impact in deep learning and other machine learning applications. Recently, quantum annealers have been proposed as a potential candidate to speed up this task, but several limitations still bar these state-of-the-art technologies from being used effectively. One of the main limitations is that, while the device may indeed sample from a Boltzmann-like distribution, quantum dynamical arguments suggests it will do so with an instance-dependent effective temperature, different from the physical temperature of the device. Unless this unknown temperature can be unveiled, it might not be possible to effectively use a quantum annealer for Boltzmann sampling. In this talk, we present a strategy to overcome this challenge with a simple effective-temperature estimation algorithm. We provide a systematic study assessing the impact of the effective temperatures in the learning of a kind of restricted Boltzmann machine embedded on quantum hardware, which can serve as a building block for deep learning architectures. We also provide a comparison to k-step contrastive divergence (CD-k) with k up to 100. Although assuming a suitable fixed effective temperature also allows to outperform one step contrastive divergence (CD-1), only when using an instance-dependent effective temperature we find a performance close to that of CD-100 for the case studied here. We discuss generalizations of the algorithm to other more expressive generative models, beyond restricted Boltzmann machines.

Perdomo-Ortiz, Alejandro↗
NASA NTRS2007

Automated Knowledge Discovery From Simulators

A computational method, SimLearn, has been devised to facilitate efficient knowledge discovery from simulators. Simulators are complex computer programs used in science and engineering to model diverse phenomena such as fluid flow, gravitational interactions, coupled mechanical systems, and nuclear, chemical, and biological processes. SimLearn uses active-learning techniques to efficiently address the "landscape characterization problem." In particular, SimLearn tries to determine which regions in "input space" lead to a given output from the simulator, where "input space" refers to an abstraction of all the variables going into the simulator, e.g., initial conditions, parameters, and interaction equations. Landscape characterization can be viewed as an attempt to invert the forward mapping of the simulator and recover the inputs that produce a particular output. Given that a single simulation run can take days or weeks to complete even on a large computing cluster, SimLearn attempts to reduce costs by reducing the number of simulations needed to effect discoveries. Unlike conventional data-mining methods that are applied to static predefined datasets, SimLearn involves an iterative process in which a most informative dataset is constructed dynamically by using the simulator as an oracle. On each iteration, the algorithm models the knowledge it has gained through previous simulation trials and then chooses which simulation trials to run next. Running these trials through the simulator produces new data in the form of input-output pairs. The overall process is embodied in an algorithm that combines support vector machines (SVMs) with active learning. SVMs use learning from examples (the examples are the input-output pairs generated by running the simulator) and a principle called maximum margin to derive predictors that generalize well to new inputs. In SimLearn, the SVM plays the role of modeling the knowledge that has been gained through previous simulation trials. Active learning is used to determine which new input points would be most informative if their output were known. The selected input points are run through the simulator to generate new information that can be used to refine the SVM. The process is then repeated. SimLearn carefully balances exploration (semi-randomly searching around the input space) versus exploitation (using the current state of knowledge to conduct a tightly focused search). During each iteration, SimLearn uses not one, but an ensemble of SVMs. Each SVM in the ensemble is characterized by different hyper-parameters that control various aspects of the learned predictor - for example, whether the predictor is constrained to be very smooth (nearby points in input space lead to similar output predictions) or whether the predictor is allowed to be "bumpy." The various SVMs will have different preferences about which input points they would like to run through the simulator next. SimLearn includes a formal mechanism for balancing the ensemble SVM preferences so that a single choice can be made for the next set of trials.

Burl, Michael↗
NASA NTRSDate not supplied

Materials Informatics at NASA GRC: Machine Learning Surrogate Modeling, Data Management, and Integrated Toolsets for Establishing/Maintaining the Digital Thread

Integrated Computational Materials Engineering (ICME) has recently received widespread attention due to its promises in reducing dependence on physical testing for engineering design by relying on simulation, reducing both time and cost to market for various applications. ICME however requires validated multiscale material models, which heavily depend on available test data with full material and test pedigree, including material processing, test and measurement equipment, raw data collection, and analysis methodology and results that is findable and usable, along with integrated, efficient toolsets for effectively passing information across various length and time scales across such models. At the NASA Glenn Research Center under the Transformational Tools and Technologies Project, significant recent efforts have been directed towards establishing the required cyberinfrastructure to enable optimized ICME processes and the design of “fit-for-purpose” materials to achieve the goals outlined in the NASA Vision 2040 report. Such efforts include development of multiscale physics-based material models, which can be used to train highly efficient surrogate machine learning models, development of best practices and infrastructure for effective, traceable materials information management, and development of toolsets that integrate with physics-based codes, machine learning models, and an information management system to enable high throughput of materials data collection and analysis, establishment of digital twins and the digital thread, and automation of the ICME design process for material optimization.

Machine Learning↗
NASA NTRS2018

Systems Development, Data Mining, and Knowledge Discovery

The primary role of the Technical Integration Office is to provide technical solutions and services to different branches at KSC (Kennedy Space Center) and NASA program customers. The Technical Integration Office helps support KSC's operational needs by providing services such as digital connectivity, data center services, modelling and simulation tools, and communication video services. To learn the necessary technology and processes for my internship, I am working on two projects: learning C# (C Sharp programming language) with SQL and developing requirements for a PX (Communication and Public Engagement) inventory management system. To learn how to efficiently program with C#, my mentor assigned me to complete a sports informatics application that would let users discover facts and rules about various sports. The sports informatics application comes with search capabilities, report generating features, rule lists that users can modify, and diagrams for various sport strategies. To further build upon this project, I also developed a sport simulation game with the application. Once I begin more SQL-based projects, I will have the opportunity to learn how to manage databases and link SQL servers with C# programs. To develop requirements for the inventory management system, I have met with PX representatives and toured their storage facilities to see how they organize and store their items and equipment. I will also be meeting with representatives from the budget office to find out what information must be in a system budget report. The main components the system must have are customer request management, a search feature for items and equipment, report generation capabilities, and automated system warnings when item quantities reach or go below administrator-specified threshold levels. I have drafted questions and shall statements that will ultimately become part of the inventory management system requirements document.

Espinosa, Gabriel↗
NASA NTRS2009

Active Learning with Irrelevant Examples

An improved active learning method has been devised for training data classifiers. One example of a data classifier is the algorithm used by the United States Postal Service since the 1960s to recognize scans of handwritten digits for processing zip codes. Active learning algorithms enable rapid training with minimal investment of time on the part of human experts to provide training examples consisting of correctly classified (labeled) input data. They function by identifying which examples would be most profitable for a human expert to label. The goal is to maximize classifier accuracy while minimizing the number of examples the expert must label. Although there are several well-established methods for active learning, they may not operate well when irrelevant examples are present in the data set. That is, they may select an item for labeling that the expert simply cannot assign to any of the valid classes. In the context of classifying handwritten digits, the irrelevant items may include stray marks, smudges, and mis-scans. Querying the expert about these items results in wasted time or erroneous labels, if the expert is forced to assign the item to one of the valid classes. In contrast, the new algorithm provides a specific mechanism for avoiding querying the irrelevant items. This algorithm has two components: an active learner (which could be a conventional active learning algorithm) and a relevance classifier. The combination of these components yields a method, denoted Relevance Bias, that enables the active learner to avoid querying irrelevant data so as to increase its learning rate and efficiency when irrelevant items are present. The algorithm collects irrelevant data in a set of rejected examples, then trains the relevance classifier to distinguish between labeled (relevant) training examples and the rejected ones. The active learner combines its ranking of the items with the probability that they are relevant to yield a final decision about which item to present to the expert for labeling. Experiments on several data sets have demonstrated that the Relevance Bias approach significantly decreases the number of irrelevant items queried and also accelerates learning speed.

Wagstaff, Kiri↗
NASA NTRS2016

ROS Hexapod

As an intern project for NASA Johnson Space Center (JSC), my job was to familiarize myself and operate a Robotics Operating System (ROS). The project outcome converted existing software assets into ROS using nodes, enabling a robotic Hexapod to communicate to be functional and controlled by an existing PlayStation 3 (PS3) controller. Existing control algorithms and current libraries have no ROS capabilities within the Hexapod C++ source code when the internship started, but that has changed throughout my internship. Conversion of C++ codes to ROS enabled existing code to be compatible with ROS, and is now controlled using an existing PS3 controller. Furthermore, my job description was to design ROS messages and script programs that enabled assets to participate in the ROS ecosystem by subscribing and publishing messages. Software programming source code is written in directories using C++. Testing of software assets included compiling code within the Linux environment using a terminal. The terminal ran the code from a directory. Several problems occurred while compiling code and the code would not compile. So modifying code to where C++ can read the source code were made. Once the code was compiled and ran, the code was uploaded to Hexapod and then controlled by a PS3 controller. The project outcome has the Hexapod fully functional and compatible with ROS and operates using the PlayStation 3 controller. In addition, an open source software (IDE) Arduino board will be integrated into the ecosystem with designing circuitry on a breadboard to add additional behavior with push buttons, potentiometers and other simple elements in the electrical circuitry. Other projects with the Arduino will be a GPS module, digital clock that will run off 22 satellites to show accurate real time using a GPS signal and an internal patch antenna to communicate with satellites. In addition, this internship experience has led me to pursue myself to learn coding more efficiently and effectively to write, subscribe and publish my own source code in different programming languages. With some familiarity with software programming, it will enhance my skills in the electrical engineering field. In contrast, my experience here at JSC with the Simulation and Graphics Branch (ER7) has led me to take my coding skill to be more proficient to increase my knowledge in software programming, and also enhancing my skills in ROS. This knowledge will be taken back to my university to implement coding in a school project that will use source coding and ROS to work on the PR2 robot which is controlled by ROS software. My skills learned here will be used to integrate messages to subscribe and publish ROS messages to a PR2 robot. The PR2 robot will be controlled by an existing PS3 controller by changing C++ coding to subscribe and publish messages to ROS. Overall the skills that were obtained here will not be lost, but increased.

Davis, Kirsch↗
NASA NTRS1977

Launch processing system transition from development to operation

The Launch Processing System has been under development at Kennedy Space Center since 1973. A prototype system was developed and delivered to Marshall Space Flight Center for Solid Rocket Booster checkout in July 1976. The first production hardware arrived in late 1976. The System uses a distributed computer network for command and monitoring and is supported by a dual large scale computer system for 'off line' processing. A high level of automation is anticipated for Shuttle and Payload testing and launch operations to gain the advantages of short turnaround capability, repeatability of operations, and minimization of operations and maintenance (O&M) manpower. Learning how to efficiently apply the system is our current problem. We are searching for more effective ways to convey LPS system performance characteristics from the designer to a large number of users. Once we have done this, we can realize the advantages of LPS system design.

Paul, H. C.↗
NASA NTRS1989

Rhesus monkeys (Macaca mulatta), video tasks, and implications for stimulus-response spatial contiguity

Recent reports support the argument that the efficiency of primate learning is compromised to the degree that there is spatial discontiguity between discriminands and the locus of response. Experiments are reported here in which two rhesus monkeys easily mastered precise control of a joystick to respond to a variety of computer-generated targets despite the fact that the joystick was located 9 to 18 cm from the video screen. It is argued that stimulus-response contiguity is a significant parameter of learning only to the degree that the monkey visually attends to the directional movements of its hand in order to displace discriminands. If attention is focused on the effects of the hand's movement rather than on the hand itself, stimulus-response contiguity is no longer a primary parameter of learning. The implications of these results for mirror-guided studies are discussed.

Rumbaugh, Duane M.↗
NASA NTRS1988

Initial test results for the mini-mast

The objectives of the 20-meter Mini-Mast were: (1) to learn how to efficiently test this type of large truss structure, (2) to relate component testing to the overall behavior of the structure, and (3) to update the associated analytical model based upon the experimental data. The Mini-Mast represents structural characteristics similar to the COFS beam which is planned to be flown on Shuttle to perform on-orbit structures and controls experiments. The information is of interest to LDR because it represents analysis and test information on a truss-type structure which may be similar to the Large Deployable Reflector (LDR) backup structure. The structure has a total of 111 titanium joints; the joint in the center of the truss element is a near-center latch joint. The results of this research indicate that linear deployable-type structures can be built, but difficulties do exist in extracting modes with identical frequencies; gravitational loading does affect the ground test results; and prediction of truss-type-structure dynamic characteristics is not trivial.

Horta, L.↗
NASA NTRS1992

Adaptive defuzzification for fuzzy systems modeling

We propose a new parameterized method for the defuzzification process based on the simple M-SLIDE transformation. We develop a computationally efficient algorithm for learning the relevant parameter as well as providing a computationally simple scheme for doing the defuzzification step in the fuzzy logic controllers. The M-SLIDE method results in a particularly simple linear form of the algorithm for learning the parameter which can be used both off- and on-line.

Yager, Ronald R.↗
NASA NTRS1997

Design, Modeling and Performance Optimization of a Novel Rotary Piezoelectric Motor

This work has demonstrated a proof of concept for a torsional inchworm type motor. The prototype motor has shown that piezoelectric stack actuators can be used for rotary inchworm motor. The discrete linear motion of piezoelectric stacks can be converted into rotary stepping motion. The stacks with its high force and displacement output are suitable actuators for use in piezoelectric motor. The designed motor is capable of delivering high torque and speed. Critical issues involving the design and operation of piezoelectric motors were studied. The tolerance between the contact shoes and the rotor has proved to be very critical to the performance of the motor. Based on the prototype motor, a waveform optimization scheme was proposed and implemented to improve the performance of the motor. The motor was successfully modeled in MATLAB. The model closely represents the behavior of the prototype motor. Using the motor model, the input waveforms were successfully optimized to improve the performance of the motor in term of speed, torque, power and precision. These optimized waveforms drastically improve the speed of the motor at different frequencies and loading conditions experimentally. The optimized waveforms also increase the level of precision of the motor. The use of the optimized waveform is a break-away from the traditional use of sinusoidal and square waves as the driving signals. This waveform optimization scheme can be applied to any inchworm motors to improve their performance. The prototype motor in this dissertation as a proof of concept was designed to be robust and large. Future motor can be designed much smaller and more efficient with lessons learned from the prototype motor.

Duong, Khanh A.↗
NASA NTRS2004

Test Planning Approach and Lessons

As NASA began technology risk reduction activities and planning for the next generation launch vehicle under the Space Launch Initiative (SLI), now the Next Generation Launch Technology (NGLT) Program, a review of past large liquid rocket engine development programs was performed. The intent of the review was to identify any significant lessons from the development testing programs that could be applied to current and future engine development programs. Because the primary prototype engine in design at the time of this study was the Boeing-Rocketdyne RS-84, the study was slightly biased towards LOX/RP-1 liquid propellant engines. However, the significant lessons identified are universal. It is anticipated that these lessons will serve as a reference for test planning in the Engine Systems Group at Marshall Space Flight Center (MSFC). Towards the end of F-1 and J-2 engine development testing, NASA/MSFC asked Rocketdyne to review those test programs. The result was a document titled, Study to Accelerate Development by Test of a Rocket Engine (R-8099). The "intent (of this study) is to apply this thinking and learning to more efficiently develop rocket engines to high reliability with improved cost effectivenes" Additionally, several other engine programs were reviewed - such as SSME, NSTS, STME, MC-1, and RS-83- to support or refute the R-8099. R-8099 revealed two primary lessons for test planning, which were supported by the other engine development programs. First, engine development programs can benefit from arranging the test program for engine system testing as early as feasible. The best test for determining environments is at the system level, the closest to the operational flight environment. Secondly, the component testing, which tends to be elaborate, should instead be geared towards reducing risk to enable system test. Technical risk can be reduced at the component level, but the design can only be truly verified and validated after engine system testing.

Parkinson, Douglas A.↗
NASA NTRS2014

In-Situ Resource Utilization for Space Exploration: Resource Processing, Mission-Enabling Technologies, and Lessons for Sustainability on Earth and Beyond

As humanity begins to reach out into the solar system, it has become apparent that supporting a human or robotic presence in transit and/or on station requires significant expendable resources including consumables (to support people), fuel, and convenient reliable power. Transporting all necessary expendables is inefficient, inconvenient, costly, and, in the final analysis, a complicating factor for mission planners and a significant source of potential failure modes. Over the past twenty-five years, beginning with the Space Exploration Initiative, researchers at the NASA Glenn Research Center (GRC), academic collaborators, and industrial partners have analyzed, researched, and developed successful solutions for the challenges posed by surviving and even thriving in the resource limited environment(s) presented by near-Earth space and non-terrestrial surface operations. In this retrospective paper, we highlight the efforts of the co-authors in resource simulation and utilization, materials processing and consumable(s) production, power systems and analysis, fuel storage and handling, propulsion systems, and mission operations. As we move forward in our quest to explore space using a resource-optimized approach, it is worthwhile to consider lessons learned relative to efficient utilization of the (comparatively) abundant natural resources and improving the sustainability (and environment) for life on Earth. We reconsider Lunar (and briefly Martian) resource utilization for potential colonization, and discuss next steps moving away from Earth.

Hepp, A. F.↗
NASA NTRS2015

In-Situ Resource Utilization for Space Exploration: Resource Processing, Mission-Enabling Technologies, and Lessons for Sustainability on Earth and Beyond

As humanity begins to reach out into the solar system, it has become apparent that supporting a human or robotic presence in transit andor on station requires significant expendable resources including consumables (to support people), fuel, and convenient reliable power. Transporting all necessary expendables is inefficient, inconvenient, costly, and, in the final analysis, a complicating factor for mission planners and a significant source of potential failure modes. Over the past twenty-five years, beginning with the Space Exploration Initiative, researchers at the NASA Glenn Research Center (GRC), academic collaborators, and industrial partners have analyzed, researched, and developed successful solutions for the challenges posed by surviving and even thriving in the resource limited environment(s) presented by near-Earth space and non-terrestrial surface operations. In this retrospective paper, we highlight the efforts of the co-authors in resource simulation and utilization, materials processing and consumable(s) production, power systems and analysis, fuel storage and handling, propulsion systems, and mission operations. As we move forward in our quest to explore space using a resource-optimized approach, it is worthwhile to consider lessons learned relative to efficient utilization of the (comparatively) abundant natural resources and improving the sustainability (and environment) for life on Earth. We reconsider Lunar (and briefly Martian) resource utilization for potential colonization, and discuss next steps moving away from Earth.

Specific impulse↗
NASA NTRSDate not supplied

Statistical learning framework for safety and failure analysis of a DNN-based autonomous aircraft system

Deep Neural Networks (DNNs) and Machine Learning technology is increasingly used for safety-critical applications in the Aerospace domain. To ensure safe operations, the DNN and the system must undergo rigorous verification and validation, including advanced statistical analyses. Performance and safety of the DNN and system behavior must not only be analyzed for the nominal case, but under numerous off-nominal and failure cases. In this paper we will describe how our statistical learning framework SYSAI can efficiently perform such analyses using the tool’s unique combination of advanced learning modeling and statistical analysis techniques. SYSAI can effectively explore the high-dimensional state and failure space of the system under test; geometrical shape detection of safety regions and boundaries support explainability of the results to the designer. In this paper, we report experiments and results obtained with a vision-based DNN control system (ACT) that is capable of autonomously steering an aircraft down a runway.

Yuning He↗