Search NASA⌕ Search

SEARCH · Search NASA

Results for “Learning systems”

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 1,279 records · Page 71
All sourcesNASA NTRSDOE OSTI
NASA NTRS1989

Production system chunking in SOAR: Case studies in automated learning

A preliminary study of SOAR, a general intelligent architecture for automated problem solving and learning, is presented. The underlying principles of universal subgoaling and chunking were applied to a simple, yet representative, problem in artificial intelligence. A number of problem space representations were examined and compared. It is concluded that learning is an inherent and beneficial aspect of problem solving. Additional studies are suggested in domains relevant to mission planning and to SOAR itself.

Allen, Robert↗
NASA NTRS1991

IVA the robot: Design guidelines and lessons learned from the first space station laboratory manipulation system

The first interactive Space Station Freedom (SSF) lab robot exhibit was installed at the Space and Rocket Center in Huntsville, AL, and has been running daily since. IntraVehicular Activity (IVA) the robot is mounted in a full scale U.S. Lab (USL) mockup to educate the public on possible automation and robotic applications aboard the SSF. Responding to audio and video instructions at the Command Console, exhibit patrons may prompt IVA to perform a housekeeping task or give a speaking tour of the module. Other exemplary space station tasks are simulated and the public can even challenge IVA to a game of tic tac toe. In anticipation of such a system being built for the Space Station, a discussion is provided of the approach taken, along with suggestions for applicability to the Space Station Environment.

Konkel, Carl R.↗
NASA NTRS1990

Introduction to Fuzzy Set Theory

An introduction to fuzzy set theory is described. Topics covered include: neural networks and fuzzy systems; the dynamical systems approach to machine intelligence; intelligent behavior as adaptive model-free estimation; fuzziness versus probability; fuzzy sets; the entropy-subsethood theorem; adaptive fuzzy systems for backing up a truck-and-trailer; product-space clustering with differential competitive learning; and adaptive fuzzy system for target tracking.

Kosko, Bart↗
NASA NTRS1993

Experience, lessons learned and methodology in the design of space systems to accommodate total dose and SEU effects

We now have considerable experience with successfully designing science systems to function properly, even during and after exposure to ionizing radiation approaching one Megarad, and during large solar particle events with substantial high Z fluxes such as the events that happened in August, 1972 and in the fall of 1989. Nevertheless, with changing device dimensions and properties, newer technologies and new applications, new problems arise. Some recent basic research has shed light on these problems and pointed the way to better diagnostic tests. In the case of single event upsets, clearly one cannot always accurately predict upset probabilities simply by knowing the quiescent Linear Energy Transfer (LET) threshold and the asymptotic cross-section of the device.

Trainor, James H.↗
NASA NTRS2010

Space Shuttle Program Primary Avionics Software System (PASS) Success Legacy -Major Accomplishments and Lessons Learned

This presentation has shown the accomplishments of the PASS project over three decades and highlighted the lessons learned. Over the entire time, our goal has been to continuously improve our process, implement automation for both quality and increased productivity, and identify and remove all defects due to prior execution of a flawed process in addition to improving our processes following identification of significant process escapes. Morale and workforce instability have been issues, most significantly during 1993 to 1998 (period of consolidation in aerospace industry). The PASS project has also consulted with others, including the Software Engineering Institute, so as to be an early evaluator, adopter, and adapter of state-of-the-art software engineering innovations.

Orr, James K.↗
NASA NTRSDate not supplied

Challenges, Lessons Learned, and Methodologies from the LCRD Optical Communication System AI&T

The Laser Communications Relay Demonstration (LCRD) is a space flight technology demonstration mission, led by the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center (GSFC) in Greenbelt, Maryland and sponsored by NASA’s Technology Demonstration Missions (TDM) Program and Space Communications and Navigation (SCaN) Program Office. The LCRD payload is hosted on the Department of Defense (DoD) Space Test Program (STP) Satellite-6 (STPSat-6) space vehicle and will operate in geostationary orbit (GEO). Launching in late 2021, the mission will conduct a minimum of two years of communication experiments with optical terminals at NASA’s Jet Propulsion Laboratory (JPL) Table Mountain Facility, in Hawaii, on the International Space Station in LEO, and via a high bandwidth radio link to White Sands Complex (WSC), New Mexico. This paper focuses on the assembly, integration, and test (AI&T) campaign spanning more than four years, using multiple test facilities, and involving multiple partner collaborations.

Bernie Edwards↗
DOE OSTI2020

The Single-Volume Scatter Camera

The multi-institution Single-Volume Scatter Camera (SVSC) collaboration led by Sandia National Laboratories (SNL) is developing a compact, high-efficiency double-scatter neutron imaging system. Kinematic emission imaging of fission-energy neutrons can be used to detect, locate, and spatially characterize special nuclear material. Neutron-scatter cameras, analogous to Compton imagers for gamma ray detection, have a wide field of view, good event-by-event angular resolution, and spectral sensitivity. Existing systems, however, suffer from large size and/or poor efficiency. We are developing high-efficiency scatter cameras with small form factors by detecting both neutron scatters in a compact active volume. This effort requires development and characterization of individual system components, namely fast organic scintillators, photodetectors, electronics, and reconstruction algorithms. In this presentation, we will focus on characterization measurements of several SVSC candidate scintillators. The SVSC collaboration is investigating two system concepts: the monolithic design in which isotropically emitted photons are detected on the sides of the volume, and the optically segmented design in which scintillation light is channeled along scintillator bars to segmented photodetector readout. For each of these approaches, we will describe the construction and performance of prototype systems. We will conclude by summarizing lessons learned, comparing and contrasting the two system designs, and outlining plans for the next iteration of prototype design and construction.

Manfredi, Juan↗
DOE OSTI2020

Enforcing constraints for time series prediction in supervised, unsupervised and reinforcement learning

We assume that we are given a time series of data from a dynamical system and our task is to learn the flow map of the dynamical system. We present a collection of results on how to enforce constraints coming from the dynamical system in order to accelerate the training of deep neural networks to represent the flow map of the system as well as increase their predictive ability. In particular, we provide ways to enforce constraints during training for all three major modes of learning, namely supervised, unsupervised and reinforcement learning. In general, the dynamic constraints need to include terms which are analogous to memory terms in model reduction formalisms. Such memory terms act as a restoring force which corrects the errors committed by the learned flow map during prediction. For supervised learning, the constraints are added to the objective function. For the case of unsupervised learning, in particular generative adversarial networks, the constraints are introduced by augmenting the input of the discriminator. Finally, for the case of reinforcement learning and in particular actor-critic methods, the constraints are added to the reward function. In addition, for the reinforcement learning case, we present a novel approach based on homotopy of the action-value function in order to stabilize and accelerate training. We use numerical results for the Lorenz system to illustrate the various constructions.

Stinis, Panagiotis↗