Search NASA⌕ Search

SEARCH · Search NASA

Results for “Generalization error”

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 577 records · Page 32
All sourcesNASA NTRSDOE OSTI
NASA NTRS1978

Directional errors of movements and their correction in a discrete tracking task

Subjects can correct their own errors of movement more quickly than they can react to external stimuli by using three general categories of feedback: (1) knowledge of results, primarily visually mediated; (2) proprioceptive or kinaesthetic such as from muscle spindles and joint receptors, and (3) corollary discharge or efference copy within the central nervous system. The effects of these feedbacks on simple reaction time, choice reaction time, and error correction time were studied in four normal human subjects. The movement used was plantarflexion and dorsiflexion of the ankle joint. The feedback loops were modified, by changing the sign of the visual display to alter the subject's perception of results, and by applying vibration at 100 Hz simultaneously to both the agonist and antagonist muscles of the ankle joint. The central processing was interfered with when the subjects were given moderate doses of alcohol (blood alcohol concentration levels of up to 0.07%). Vibration and alcohol increase both the simple and choice reaction times but not the error correction time.

Jaeger, R. J.↗
NASA NTRS1994

Global and regional kinematics with GPS

The inherent precision of the doubly differenced phase measurement and the low cost of instrumentation made GPS the space geodetic technique of choice for regional surveys as soon as the constellation reached acceptable geometry in the area of interest: 1985 in western North America, the early 1990's in most of the world. Instrument and site-related errors for horizontal positioning are usually less than 3 mm, so that the dominant source of error is uncertainty in the reference frame defined by the satellites orbits and the tracking stations used to determine them. Prior to about 1992, when the tracking network for most experiments was globally sparse, the number of fiducial sites or the level at which they could be tied to an SLR or VLBI reference frame usually, set the accuracy limit. Recently, with a global network of over 30 stations, the limit is set more often by deficiencies in models for non-gravitational forces acting on the satellites. For regional networks in the northern hemisphere, reference frame errors are currently about 3 parts per billion (ppb) in horizontal position, allowing centimeter-level accuracies over intercontinental distances and less than 1 mm for a 100 km baseline. The accuracy of GPS measurements for monitoring height variations is generally 2-3 times worse than for horizontal motions. As for VLBI, the primary source of error is unmodeled fluctuations in atmospheric water vapor, but both reference frame uncertainties and some instrument errors are more serious for vertical than horizontal measurements. Under good conditions, daily repeatabilities at the level of 10 mm rms were achieved. This paper will summarize the current accuracy of GPS measurements and their implication for the use of SLR to study regional kinematics.

King, Robert W.↗
NASA NTRS1988

Introduction to a system for implementing neural net connections on SIMD architectures

Neural networks have attracted much interest recently, and using parallel architectures to simulate neural networks is a natural and necessary application. The SIMD model of parallel computation is chosen, because systems of this type can be built with large numbers of processing elements. However, such systems are not naturally suited to generalized communication. A method is proposed that allows an implementation of neural network connections on massively parallel SIMD architectures. The key to this system is an algorithm permitting the formation of arbitrary connections between the neurons. A feature is the ability to add new connections quickly. It also has error recovery ability and is robust over a variety of network topologies. Simulations of the general connection system, and its implementation on the Connection Machine, indicate that the time and space requirements are proportional to the product of the average number of connections per neuron and the diameter of the interconnection network.

Tomboulian, Sherryl↗
NASA NTRS1988

Introduction to a system for implementing neural net connections on SIMD architectures

Neural networks have attracted much interest recently, and using parallel architectures to simulate neural networks is a natural and necessary application. The SIMD model of parallel computation is chosen, because systems of this type can be built with large numbers of processing elements. However, such systems are not naturally suited to generalized elements. A method is proposed that allows an implementation of neural network connections on massively parallel SIMD architectures. The key to this system is an algorithm permitting the formation of arbitrary connections between the neurons. A feature is the ability to add new connections quickly. It also has error recovery ability and is robust over a variety of network topologies. Simulations of the general connection system, and its implementation on the Connection Machine, indicate that the time and space requirements are proportional to the product of the average number of connections per neuron and the diameter of the interconnection network.

Tomboulian, Sherryl↗
NASA NTRS1989

Error analysis for semi-analytic displacement derivatives with respect to shape and sizing variables

Sensitivity analysis is fundamental to the solution of structural optimization problems. Consequently, much research has focused on the efficient computation of static displacement derivatives. As originally developed, these methods relied on analytical representations for the derivatives of the structural stiffness matrix (K) with respect to the design variables (b sub i). To extend these methods for use with complex finite element formulations and facilitate their implementation into structural optimization programs using the general finite element method analysis codes, the semi-analytic method was developed. In this method the matrix the derivative of K/the derivative b sub i is approximated by finite difference. Although it is well known that the accuracy of the semi-analytic method is dependent on the finite difference parameter, recent work has suggested that more fundamental inaccuracies exist in the method when used for shape optimization. Another study has argued qualitatively that these errors are related to nonuniform errors in the stiffness matrix derivatives. The accuracy of the semi-analytic method is investigated. A general framework was developed for the error analysis and then it is shown analytically that the errors in the method are entirely accounted for by errors in delta K/delta b sub i. Furthermore, it is demonstrated that acceptable accuracy in the derivatives can be obtained through careful selection of the finite difference parameter.

Fenyes, Peter A.↗
NASA NTRS1986

Generalized parity relations for large space structures with uncertain parameters

The generalized parity relations method is a technique that can be used to detect sensor and actuator failures on a large space structure. A model of a grid structure was used to evaluate the performance of these relations. It shows their relative sensitivity to modeling errors. A method using sensor outputs and actuator inputs is required for the design of the generalized parity relations. Three different estimators are studied. The last estimator can generate relations optimized for the detection of a particular failure which are interesting when the level of sensor noise is high.

Dutilloy, J. R.↗
NASA NTRS2021

Quantifying the Observational Requirements of a Space-borne LiDAR Snow Mission

This study quantifies the level of observational accuracy required from a spaceborne light detection and ranging (LiDAR) snow depth retrieval mission for enabling beneficial impacts for snow estimation. The study is conducted over a region in Western Colorado using a suite of observing system simulation experiments(OSSEs).The Joint UK Land Environment Simulator, version 5.0 (JULES v5.0) is employed to simulate a suite of idealized LiDAR observations, considering a range of LiDAR snow depth retrieval errors, different hypothetical sensor swath widths, and the impact of cloud cover on observability. These simulated observations are then assimilated into the Noah land surface modelwith multi-parameterization options, version 3.6 (Noah-MP v3.6) model. This data assimilation setup is used to systematically evaluate the potential utility of LiDAR observations for improving modeled snow water equivalent (SWE)estimates and water budget variables such as runoff. Results from the OSSE runsshow that, in general, assimilation of synthetic LiDAR observations provide beneficial impacts when theLiDAR snow depth retrieval error standard deviation (σerror) is below 60 cm.Based on comparisons between the realistic (i.e., swath-limited and cloud-attenuated) case and the idealized (i.e., infinite swath width in the absence of cloud cover) case,this study concludes that observations with a conservative error standard deviation threshold of 40 cm (i.e., upper limit of the snow depth retrieval error that adds value to the SWE estimates viaassimilation) are needed for improving modeled snow estimates. More than a 33% reduction in SWEroot mean square errors and more than a 15% increase in correlation coefficientsare achieved when σ error ≤ 40 cmusing a 170-km sensor swath width in the presence of cloud attenuation effects. Further, the integrated hydrologic response, as represented by total (surface and subsurface) runoff estimates during the snow ablation season, are also enhanced when assimilating synthetic LiDAR snow depth retrievals with errors below this level.

Yonghwan Kwon↗
NASA NTRS1973

The doublet-ratio method and interstellar abundances.

The method of doublet ratios, used in obtaining interstellar Na I and Ca II abundances, is generalized to include realistic multiple-cloud cases. Entirely apart from any errors of observation, the simplified velocity distribution used in the method leads to errors in the inferred column densities which are systematic and which can be as large as a factor of ten or more in some practical cases. The D lines of Na I toward zeta Oph illustrate such order-of-magnitude underestimates.

Nachman, P.↗
NASA NTRS1973

On-line diagnosis of sequential systems

A model for on-line diagnosis was investigated for discrete-time systems, and resettable sequential systems. Generalized notions of a realization are discussed along with fault tolerance and errors. Further investigation into the theory of on-line diagnosis is recommended for three levels: binary state-assigned level, logical circuit level, and the subsystem-network level.

Sundstrom, R. J.↗
NASA NTRS1977

Optimal estimation for discrete time jump processes

Optimum estimates of nonobservable random variables or random processes which influence the rate functions of a discrete time jump process (DTJP) are obtained. The approach is based on the a posteriori probability of a nonobservable event expressed in terms of the a priori probability of that event and of the sample function probability of the DTJP. A general representation for optimum estimates and recursive equations for minimum mean squared error (MMSE) estimates are obtained. MMSE estimates are nonlinear functions of the observations. The problem of estimating the rate of a DTJP when the rate is a random variable with a probability density function of the form cx super K (l-x) super m and show that the MMSE estimates are linear in this case. This class of density functions explains why there are insignificant differences between optimum unconstrained and linear MMSE estimates in a variety of problems.

Vaca, M. V.↗
NASA NTRS1983

Comparison of simulator fidelity model predictions with in-simulator evaluation data

A full factorial in simulator experiment of a single axis, multiloop, compensatory pitch tracking task is described. The experiment was conducted to provide data to validate extensions to an analytic, closed loop model of a real time digital simulation facility. The results of the experiment encompassing various simulation fidelity factors, such as visual delay, digital integration algorithms, computer iteration rates, control loading bandwidths and proprioceptive cues, and g-seat kinesthetic cues, are compared with predictions obtained from the analytic model incorporating an optimal control model of the human pilot. The in-simulator results demonstrate more sensitivity to the g-seat and to the control loader conditions than were predicted by the model. However, the model predictions are generally upheld, although the predicted magnitudes of the states and of the error terms are sometimes off considerably. Of particular concern is the large sensitivity difference for one control loader condition, as well as the model/in-simulator mismatch in the magnitude of the plant states when the other states match.

Parrish, R. V.↗
NASA NTRS1983

New Syndrome Decoding Techniques for the (n, K) Convolutional Codes

This paper presents a new syndrome decoding algorithm for the (n,k) convolutional codes (CC) which differs completely from an earlier syndrome decoding algorithm of Schalkwijk and Vinck. The new algorithm is based on the general solution of the syndrome equation, a linear Diophantine equation for the error polynomial vector E(D). The set of Diophantine solutions is a coset of the CC. In this error coset a recursive, Viterbi-like algorithm is developed to find the minimum weight error vector (circumflex)E(D). An example, illustrating the new decoding algorithm, is given for the binary nonsystemmatic (3,1)CC.

Reed, I. S.↗
NASA NTRS1984

New syndrome decoding techniques for the (n, k) convolutional codes

This paper presents a new syndrome decoding algorithm for the (n, k) convolutional codes (CC) which differs completely from an earlier syndrome decoding algorithm of Schalkwijk and Vinck. The new algorithm is based on the general solution of the syndrome equation, a linear Diophantine equation for the error polynomial vector E(D). The set of Diophantine solutions is a coset of the CC. In this error coset a recursive, Viterbi-like algorithm is developed to find the minimum weight error vector (circumflex)E(D). An example, illustrating the new decoding algorithm, is given for the binary nonsystemmatic (3, 1)CC. Previously announced in STAR as N83-34964

Reed, I. S.↗
NASA NTRS1984

Forecasts of the 500 mb height using a dynamically oriented statistical model

The forecast skill of a simple dynamically inspired statistical model of the Northern Hemisphere 500 mb height field is evaluated in spectral and physical space for a variety of forecast lead times (1-32 days) and predictand averaging times (1-32 days). The model includes viscous damping, wave propagation, climatology and implicit stochastic forcing. The largest model skill was found for forecasts of the zonal flow and the largest waves. In general, the largest forecast skills were also associated with the largest forecast error, there being a slight geographic phase shift of the skill with respect to the error. Model skills for climate (time-averaged) forecasts are greater when using instantaneous rather than time averages to forecast time averages. Analysis of model errors suggests areas for improvement in representing forcing terms and model physics. However, the model error fields are largely 'white noise' which suggests that global forecast skills substantially larger than those obtained here are unlikely to be achieved by more sophisticated models.

Roads, J. O.↗
NASA NTRS1985

Nonsimultaneous coupled laser velocimeter measurement technique: Error prediction for spatially noncoincident measurements

A technique for obtaining orthogonal velocity components from nonorthogonal measurements using the NASA Ames Research Center Long-Range Laser Velocimeter (LRLV) is briefly discussed. A description is then presented of the error that occurs when these nonorthogonal measurements are spatially noncoincident because of positioning inaccuracies, and equations are developed for predicting this error. Sample data are presented and a prediction of the expected error for two typical applications is made. To cover other cases in general, a parametric study is conducted and the results are presented in a tabular format.

Reinath, M. S.↗
NASA NTRS1986

VLSI Architectures for Computing DFT's

Simplifications result from use of residue Fermat number systems. System of finite arithmetic over residue Fermat number systems enables calculation of discrete Fourier transform (DFT) of series of complex numbers with reduced number of multiplications. Computer architectures based on approach suitable for design of very-large-scale integrated (VLSI) circuits for computing DFT's. General approach not limited to DFT's; Applicable to decoding of error-correcting codes and other transform calculations. System readily implemented in VLSI.

Truong, T. K.↗
NASA NTRS1984

A Fault Tolerant System for an Integrated Avionics Sensor Configuration

An aircraft sensor fault tolerant system methodology for the Transport Systems Research Vehicle in a Microwave Landing System (MLS) environment is described. The fault tolerant system provides reliable estimates in the presence of possible failures both in ground-based navigation aids, and in on-board flight control and inertial sensors. Sensor failures are identified by utilizing the analytic relationships between the various sensors arising from the aircraft point mass equations of motion. The estimation and failure detection performance of the software implementation (called FINDS) of the developed system was analyzed on a nonlinear digital simulation of the research aircraft. Simulation results showing the detection performance of FINDS, using a dual redundant sensor compliment, are presented for bias, hardover, null, ramp, increased noise and scale factor failures. In general, the results show that FINDS can distinguish between normal operating sensor errors and failures while providing an excellent detection speed for bias failures in the MLS, indicated airspeed, attitude and radar altimeter sensors.

Caglayan, A. K.↗
NASA NTRS1987

Application of adaptive antenna techniques to future commercial satellite communication

The purpose of this contract was to identify the application of adaptive antenna technique in future operational commercial satellite communication systems and to quantify potential benefits. The contract consisted of two major subtasks. Task 1, Assessment of Future Commercial Satellite System Requirements, was generally referred to as the Adaptive section. Task 2 dealt with Pointing Error Compensation Study for a Multiple Scanning/Fixed Spot Beam Reflector Antenna System and was referred to as the reconfigurable system. Each of these tasks was further sub-divided into smaller subtasks. It should also be noted that the reconfigurable system is usually defined as an open-loop system while the adaptive system is a closed-loop system. The differences between the open- and closed-loop systems were defined. Both the adaptive and reconfigurable systems were explained and the potential applications of such systems were presented in the context of commercial communication satellite systems.

Ersoy, L.↗