SEARCH · Search NASA
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.
We present a model of cerebellar cortex that combines two types of learning: feedforward predicitve association based on local Hebbian-type learning between granule cell ascending branch and parallel fiber inputs, and reinforcement learning with feedback error correction based on climbing fiber activity.
Binary neutron star (NS) mergers are among the most promising astrophysical sources of gravitational wave emission for Advanced LIGO and Advanced Virgo, expected to be operational in 2015 . Finding electromagnetic counterparts to these signals will be essential to placing them in an astronomical context. The Swift satellite carries a sensitive X-ray telescope (XRT), and can respond to target-of-opportunity requests within 1-2 hours, and so is uniquely poised to find the X-ray counterparts to LIGO / Virgo triggers. Assuming NS mergers are the progenitors of short gamma-ray bursts (GRBs), some percentage of LIGO/Virgo triggers will be accompanied by X-ray band afterglows that are brighter than 10(exp -12) ergs/s/sq cm in the XRT band one day after the trigger time. We find that a soft X-ray transient of this flux is bright enough to be extremely rare, and so could be confidently associated with even a moderately localized GW signal. We examine two possible search strategies with the Swift XRT to find bright transients in LIGO/Virgo error boxes. In the first strategy, XRT could search a volume of space with a approx.100 Mpc radius by observing approx 30 galaxies over the course of a day, with sufficient depth to observe the expected X-ray afterglow. For an extended LIGO / Virgo horizon distance, the XRT could employ very short 100 s exposures to cover an area of approx 35 square degrees in about a day, and still be sensitive enough to image GW discovered GRB afterglows. These strategies demonstrate that the high X-ray luminosity of short GRBs and the relatively low X-ray transient background combine to make high confidence discoveries of X-ray band counterparts to GW triggers possible, though challenging, with current satellite facilities.
The Orbiting Carbon Observatory-2 (OCO-2), scheduled to launch in July 2014, is a NASA mission designed to measure atmospheric CO2. Its main purpose is to allow inversions of net flux estimates of CO2 on regional to continental scales using the total column CO2 retrieved using high-resolution spectra in the 0.76, 1.6, and 2.0 nm ranges. Recently, it was shown that solar-induced chlorophyll fluorescence (SIF), a proxy for gross primary production (GPP, carbon uptake through photosynthesis), can be accurately retrieved from space using high spectral resolution radiances in the 750 nm range from the Japanese GOSAT and European GOME-2 instruments. Here, we use real OCO-2 thermal vacuum test data as well as a full repeat cycle (16 days) of simulated OCO-2 spectra under realistic conditions to evaluate the potential of OCO-2 for retrievals of chlorophyll fluorescence and also its dependence on clouds and aerosols. We find that the single-measurement precision is 0.3-0.5 Wm(exp -2)sr(exp -1) nm(exp -1) (15-25% of typical peak values), better than current measurements from space but still difficult to interpret on a single-sounding basis. The most significant advancement will come from smaller ground-pixel sizes and increased measurement frequency, with a 100-fold increase compared to GOSAT (and about 8 times higher than GOME-2). This will largely decrease the need for coarse spatial and temporal averaging in data analysis and pave the way to accurate local studies.We also find that the lack of full global mapping from the OCO-2 only incurs small representativeness errors on regional averages. Eventually, the combination of net ecosystem exchange (NEE) derived from CO2 source/sink inversions and SIF as proxy for GPP from the same satellite will provide a more process-based understanding of the global carbon cycle.
We present a novel inversion algorithm that generates a mass balance field that is simultaneously consistent with independent observations of glacier inventory derived from optical imagery, cryosphere-attributed mass trends derived from satellite gravimetry, and ice surface elevation trends derived from airborne and satellite altimetry. We use this algorithm to assess mass balance across Greenland and the Canadian Arctic over the Sep-2003 to Oct- 2009 period at 26 kilometers resolution. We evaluate local algorithm-inferred mass balance against forty in situ point observations. This evaluation yields a root mean squared error (RMSE) of 0.15 mWE/a ( 0.15 meters (water equivalent) per annum), and highlights a paucity of in situ observations from regions of high dynamic mass loss and peripheral glaciers. We assess mass losses of 212 plus or minus 67 Gigatons per annum to the Greenland ice sheet proper, 38 Gigatons per annum to peripheral glaciers in Greenland, and 42 Gigatons per annum to glaciers in the Canadian Arctic. These magnitudes of mass loss are dependent on the gravimetry-derived spherical harmonic mass trend we invert. We spatially partition the transient glacier continuity equation by differencing algorithm inferred mass balance from modeled surface mass balance, in order to solve the horizontal divergence of ice flux as a residual. This residual ice dynamic field infers flux divergence (or submergent flow) in the ice sheet accumulation area and at tidewater margins, and flux convergence (or emergent flow) in land-terminating ablation areas, which is consistent with continuum mechanics theory.
We present a source inversion technique for chemical constituents that uses assimilated constituent observations rather than directly using the observations. The method is tested with a simple model problem, which is a two-dimensional Fourier-Galerkin transport model combined with a Kalman filter for data assimilation. Inversion is carried out using a Green's function method and observations are simulated from a true state with added Gaussian noise. The forecast state uses the same spectral spectral model, but differs by an unbiased Gaussian model error, and emissions models with constant errors. The numerical experiments employ both simulated in situ and satellite observation networks. Source inversion was carried out by either direct use of synthetically generated observations with added noise, or by first assimilating the observations and using the analyses to extract observations. We have conducted 20 identical twin experiments for each set of source and observation configurations, and find that in the limiting cases of a very few localized observations, or an extremely large observation network there is little advantage to carrying out assimilation first. However, in intermediate observation densities, there decreases in source inversion error standard deviation using the Kalman filter algorithm followed by Green's function inversion by 50% to 95%.
This innovation can be run as a standalone network application on any computer in a networked environment. This design can be configured to control one or more routers (one instance per router), and can also be configured to listen to a policy server over the network to receive new policies based on the policy- based network management technology. The Router Agent Technology transforms the received policies into suitable Access Control List syntax for the routers it is configured to control. It commits the newly generated access control lists to the routers and provides feedback regarding any errors that were faced. The innovation also automatically generates a time-stamped log file regarding all updates to the router it is configured to control. This technology, once installed on a local network computer and started, is autonomous because it has the capability to keep listening to new policies from the policy server, transforming those policies to router-compliant access lists, and committing those access lists to a specified interface on the specified router on the network with any error feedback regarding commitment process. The stand-alone application is named RouterAgent and is currently realized as a fully functional (version 1) implementation for the Windows operating system and for CISCO routers.
In a practical robotic system, it is important to represent and plan sequences of operations and to be able to choose an efficient sequence from them for a specific task. During the generation and execution of task plans, different kinds of uncertainty may occur and erroneous states need to be handled to ensure the efficiency and reliability of the system. An approach to task representation, planning, and error recovery for robotic systems is demonstrated. Our approach to task planning is based on an AND/OR net representation, which is then mapped to a Petri net representation of all feasible geometric states and associated feasibility criteria for net transitions. Task decomposition of robotic assembly plans based on this representation is performed on the Petri net for robotic assembly tasks, and the inheritance of properties of liveness, safeness, and reversibility at all levels of decomposition are explored. This approach provides a framework for robust execution of tasks through the properties of traceability and viability. Uncertainty in robotic systems are modeled by local fuzzy variables, fuzzy marking variables, and global fuzzy variables which are incorporated in fuzzy Petri nets. Analysis of properties and reasoning about uncertainty are investigated using fuzzy reasoning structures built into the net. Two applications of fuzzy Petri nets, robot task sequence planning and sensor-based error recovery, are explored. In the first application, the search space for feasible and complete task sequences with correct precedence relationships is reduced via the use of global fuzzy variables in reasoning about subgoals. In the second application, sensory verification operations are modeled by mutually exclusive transitions to reason about local and global fuzzy variables on-line and automatically select a retry or an alternative error recovery sequence when errors occur. Task sequencing and task execution with error recovery capability for one and multiple soft components in robotic systems are investigated.
Future lunar landing systems, particularly those used to land humans on the lunar surface as part of the ARTEMIS program, will require precision navigation relative to the lunar surface. The most common way to meet these stringent navigation requirements is through terrain relative navigation (TRN), which localizes a spacecraft by comparing descent imagery with a predefined map of the surface. The accuracy achievable using TRN is limited by the accuracy of the reference Digital Elevation Map (DEM). It is therefore critical for future lunar missions that potential errors in DEMs be quantified. This paper describes one of NASA’s current efforts to develop a process for evaluating lunar DEM quality.
A reference architecture is defined for an object-oriented implementation of domains, arrays, and distributions written in the programming language Chapel. This technology primarily addresses domains that contain arrays that have regular index sets with the low-level implementation details being beyond the scope of this discussion. What is defined is a complete set of object-oriented operators that allows one to perform data distributions for domain arrays involving regular arithmetic index sets. What is unique is that these operators allow for the arbitrary regions of the arrays to be fragmented and distributed across multiple processors with a single point of access giving the programmer the illusion that all the elements are collocated on a single processor. Today's massively parallel High Productivity Computing Systems (HPCS) are characterized by a modular structure, with a large number of processing and memory units connected by a high-speed network. Locality of access as well as load balancing are primary concerns in these systems that are typically used for high-performance scientific computation. Data distributions address these issues by providing a range of methods for spreading large data sets across the components of a system. Over the past two decades, many languages, systems, tools, and libraries have been developed for the support of distributions. Since the performance of data parallel applications is directly influenced by the distribution strategy, users often resort to low-level programming models that allow fine-tuning of the distribution aspects affecting performance, but, at the same time, are tedious and error-prone. This technology presents a reusable design of a data-distribution framework for data parallel high-performance applications. Distributions are a means to express locality in systems composed of large numbers of processor and memory components connected by a network. Since distributions have a great effect on the performance of applications, it is important that the distribution strategy is flexible, so its behavior can change depending on the needs of the application. At the same time, high productivity concerns require that the user be shielded from error-prone, tedious details such as communication and synchronization.
Image Display and Analysis Systems (MIDAS) developed at NASA/Ames for the analysis of Landsat MSS images is described. The MIDAS computer power and memory, graphics, resource-sharing, expansion and upgrade, environment and maintenance, and software/user-interface requirements are outlined; the implementation hardware (including 32-bit microprocessor, 512K error-correcting RAM, 70 or 140-Mbyte formatted disk drive, 512 x 512 x 24 color frame buffer, and local-area-network transceiver) and applications software (ELAS, CIE, and P-EDITOR) are characterized; and implementation problems, performance data, and costs are examined. Planned improvements in MIDAS hardware and design goals and areas of exploration for MIDAS software are discussed.
An electronic embodiment of a neural network based associative memory in the form of a binary connection matrix is described. The nature of false memory errors, their effect on the information storage capacity of binary connection matrix memories, and a novel technique to eliminate such errors with the help of asymmetrical extra connections are discussed. The stability of the matrix memory system incorporating a unique local inhibition scheme is analyzed in terms of local minimization of an energy function. The memory's stability, dynamic behavior, and recall capability are investigated using a 32-'neuron' electronic neural network memory with a 1024-programmable binary connection matrix.
Data insufficiency, poorly conditioned matrices and singularities in equations occur regularly in complex optimization, correlation, and interdisciplinary model studies. This work concerns itself with two methods of obtaining certain physically realistic solutions to ill-conditioned or singular algebraic systems of linear equations arising from such studies. Two efficient computational solution procedures that generally lead to locally unique solutions are presented when there is insufficient data to completely define the model, or a least-squares error formulation of this system results in an ill-conditioned system of equations. If it is assumed that a reasonable estimate of the uncertain data is available in both cases cited above, then we shall show how to obtain realistic solutions efficiently, in spite of the insufficiency of independent data. The proposed methods of solution are more efficient than singular-value decomposition for dealing with such systems, since they do not require solutions for all the non-zero eigenvalues of the coefficient matrix.
A design for a natural language database interface system, called the Deductively Augmented NASA Management Decision support System (DANMDS), is presented. The DANMDS system components have been chosen on the basis of the following considerations: maximal employment of the existing NASA IBM-PC computers and supporting software; local structuring and storing of external data via the entity-relationship model; a natural easy-to-use error-free database query language; user ability to alter query language vocabulary and data analysis heuristic; and significant artificial intelligence data analysis heuristic techniques that allow the system to become progressively and automatically more useful.
Acoustic waves propagate in thin plates as guided or Lamb modes. The velocities of these modes are dispersive in that they depend not only on the material elastic properties and density, but also on the frequency. Accurate characterization of Lamb wave dispersion is important in many acoustic based nondestructive evaluation techniques. It is necessary for ultrasonic measurements in thin plates to determine elastic properties and for flaw detection and localization. In acoustic emission (AE) testing, if not taken into account, highly dispersive Lamb mode propagation can lead to large errors in source location. In this study, the pseudo Wigner-Ville distribution (PWVD) was used for measurement of group velocity dispersion of Lamb waves in a unidirectional graphite/epoxy (AS4/3502) laminate. The PWVD is one of a number of transforms which provide a time-frequency representation of a digitized time series. Broad band acoustic waves were generated by a pencil lead fracture (Hsu-Neilsen source) and were detected with broad band ultrasonic transducers. The arrival times for the lowest order symmetric (S(sub 0)) and antisymmetric (A(sub 0)) Lamb modes were determined from measurements of the time at which the respective peak amplitudes occurred in the PWVD. Measurements were made at several source-to-detector distances and a least squares fit used to calculate the velocity. Results are presented for propagation along, and perpendicular to, the fiber direction. Theoretical dispersion curves were also calculated and a comparison between theory and experiment demonstrates good agreement.
Conventional training methods for neural networks involve starting al a random location in the solution space of the network weights, navigating an error hyper surface to reach a minimum, and sometime stochastic based techniques (e.g., genetic algorithms) to avoid entrapment in a local minimum. It is further typically necessary to preprocess the data (e.g., normalization) to keep the training algorithm on course. Conversely, Bayesian based learning is an epistemological approach concerned with formally updating the plausibility of competing candidate hypotheses thereby obtaining a posterior distribution for the network weights conditioned on the available data and a prior distribution. In this paper, we developed a powerful methodology for estimating the full residual uncertainty in network weights and therefore network predictions by using a modified Jeffery's prior combined with a Metropolis Markov Chain Monte Carlo method.
Hybrid pulse-burst Doppler global velocimetry (PB-DGV)/pulse-burst, cross-correlation DGV (PB-CC-DGV) was demonstrated for the first time in the NASA Langley 4-foot Supersonic Unitary Plan Wind Tunnel, allowing simultaneous acquisition of mean and instantaneous high-repetition-rate velocity fields. The technique was used to make planar velocity measurements across an oblique shockwave generated by a large splitter plate set at a -2° angle of attack in a Mach 2.4 flow. Sequences of more than 100 consecutive instantaneous velocity images were obtained at a rate of 100 kHz. Velocity fields from both aspects of the hybrid measurement indicated errors less than 1-percent of anticipated velocities. Likewise, all assessments of measurement precision were approximately 1.6-percent of the local velocities. Additional demonstrations in different flowfields including a subsonic axi-symmetric jet and a supersonic Mars reentry vehicle model further illustrate the utility of the new hybrid method.
Work in real-time distributed computation and control has culminated in a prototype force-reflecting telemanipulation system having a dissimilar master (cable-driven, force-reflecting hand controller) and a slave (PUMA 560 robot with custom controller), an extremely high sampling rate (1000 Hz), and a low loop computation delay (5 msec). In a series of experiments with this system and five trained test operators covering over 100 hours of teleoperation, performance was measured in a series of generic and application-driven tasks with and without force feedback, and with control shared between teleoperation and local sensor referenced control. Measurements defining task performance included 100-Hz recording of six-axis force/torque information from the slave manipulator wrist, task completion time, and visual observation of predefined task errors. The task consisted of high precision peg-in-hole insertion, electrical connectors, velcro attach-de-attach, and a twist-lock multi-pin connector. Each task was repeated three times under several operating conditions: normal bilateral telemanipulation, forward position control without force feedback, and shared control. In shared control, orientation was locally servo controlled to comply with applied torques, while translation was under operator control. All performance measures improved as capability was added along a spectrum of capabilities ranging from pure position control through force-reflecting teleoperation and shared control. Performance was optimal for the bare-handed operator.
Work during the period of this report has been in three areas: (1) pressure transducer error analysis, (2) fluctuating velocity and pressure measurements in the NASA Lewis 6-inch diameter quiet jet facility, and (3) measurement analysis. A theory was developed and experimentally verified to quantify the pressure transducer velocity interference error. The theory and supporting experimental evidence show that the errors are a function of the velocity field's turbulent structure. It is shown that near the mixing layer center the errors are negligible. Turbulent velocity and pressure measurements were made in the NASA Lewis quiet jet facility. Some preliminary results are included.