Search NASA⌕ Search

SEARCH · Search NASA

Results for “generalized least squared”

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 19 records
All sourcesNASA NTRSDOE OSTI
NASA NTRS1992

Improvement of structural models using covariance analysis and nonlinear generalized least squares

The next generation of large, flexible space structures will be too light to support their own weight, requiring a system of structural supports for ground testing. The authors have proposed multiple boundary-condition testing (MBCT), using more than one support condition to reduce uncertainties associated with the supports. MBCT would revise the mass and stiffness matrix, analytically qualifying the structure for operation in space. The same procedure is applicable to other common test conditions, such as empty/loaded tanks and subsystem/system level tests. This paper examines three techniques for constructing the covariance matrix required by nonlinear generalized least squares (NGLS) to update structural models based on modal test data. The methods range from a complicated approach used to generate the simulation data (i.e., the correct answer) to a diagonal matrix based on only two constants. The results show that NGLS is very insensitive to assumptions about the covariance matrix, suggesting that a workable NGLS procedure is possible. The examples also indicate that the multiple boundary condition procedure more accurately reduces errors than individual boundary condition tests alone.

Glaser, R. J.↗
NASA NTRS2006

Classes of Split-Plot Response Surface Designs for Equivalent Estimation

When planning an experimental investigation, we are frequently faced with factors that are difficult or time consuming to manipulate, thereby making complete randomization impractical. A split-plot structure differentiates between the experimental units associated with these hard-to-change factors and others that are relatively easy-to-change and provides an efficient strategy that integrates the restrictions imposed by the experimental apparatus. Several industrial and scientific examples are presented to illustrate design considerations encountered in the restricted randomization context. In this paper, we propose classes of split-plot response designs that provide an intuitive and natural extension from the completely randomized context. For these designs, the ordinary least squares estimates of the model are equivalent to the generalized least squares estimates. This property provides best linear unbiased estimators and simplifies model estimation. The design conditions that allow for equivalent estimation are presented enabling design construction strategies to transform completely randomized Box-Behnken, equiradial, and small composite designs into a split-plot structure.

Parker, Peter A.↗
NASA NTRS1972

Statistical assessment of a unique time series analysis technique

Empirical tests of a particular least squares multiple prediction program were studied, as well as the relationship of this particular program to general least squares multiple prediction theory. Empirical evaluation and test of the program involved: (1) conversion of the program to run on an IBM 360/44 with replication of test results obtained on a Univac 1108; and (2) generation and analysis of Monte Carlo simulated data with the objective of comparison against results theoretically obtainable from spectrum analysis routines.

Benignus, V. A.↗
NASA NTRS1988

Linear Least Squares for Correlated Data

Throughout the literature authors have consistently discussed the suspicion that regression results were less than satisfactory when the independent variables were correlated. Camm, Gulledge, and Womer, and Womer and Marcotte provide excellent applied examples of these concerns. Many authors have obtained partial solutions for this problem as discussed by Womer and Marcotte and Wonnacott and Wonnacott, which result in generalized least squares algorithms to solve restrictive cases. This paper presents a simple but relatively general multivariate method for obtaining linear least squares coefficients which are free of the statistical distortion created by correlated independent variables.

Dean, Edwin B.↗
NASA NTRS1975

The significance of the Skylab altimeter experiment results and potential applications

The Skylab Altimeter Experiment has proven the capability of the altimeter for measurement of sea surface topography. The geometric determination of the geoid/mean sea level from satellite altimetry is a new approach having significant applications in many disciplines including geodesy and oceanography. A Generalized Least Squares Collocation Technique was developed for determination of the geoid from altimetry data. The technique solves for the altimetry geoid and determines one bias term for the combined effect of sea state, orbit, tides, geoid, and instrument error using sparse ground truth data. The influence of errors in orbit and a priori geoid values are discussed. Although the Skylab altimeter instrument accuracy is about + or - 1 m, significant results were obtained in identification of large geoidal features such as over the Puerto Rico trench. Comparison of the results of several passes shows that good agreement exists between the general slopes of the altimeter geoid and the ground truth, and that the altimeter appears to be capable of providing more details than are now available with best known geoids. The altimetry geoidal profiles show excellent correlations with bathymetry and gravity. Potential applications of altimetry results to geodesy, oceanography, and geophysics are discussed.

Mourad, A. G.↗
NASA NTRS1975

Calibration and evaluation of Skylab altimetry for geodetic determination of the geoid

The author has identified the following significant results. The Skylab altimeter experiment has proven the capability of the altimeter for measurement of sea surface topography. The geometric determination of the geoid/mean sea level from satellite altimetry is a new approach having significant applications in many disciplines including geodesy and oceanography. A generalized least squares collocation technique was developed for determination of the geoid from altimetry data. The technique solves for the altimetry geoid and determines one bias term for the combined effect of sea state, orbit, tides, geoid, and instrument error using sparse ground truth data. The influence of errors in orbit and a priori geoid values are discussed. Although the Skylab altimeter instrument accuracy is about plus or minus 1m, significant results were obtained in identification of large geoidal features such as over the Puerto Rico trench. Comparison of the results of several passes shows that good agreement exists between the general slopes of the altimeter geoid and the ground truth, and that the altimeter appears to be capable of providing more details than are now available with best known geoids.

Mourad, A. G.↗
NASA NTRS1973

A generalized technique for using cones and dihedral angles in attitude determination, revision 1

Analytic development is presented for a general least squares attitude determination subroutine applicable to spinning satellites. The method is founded on a geometric approach which is completely divorced from considerations relating to particular types and configurations of onboard attitude sensors. Any mix of sensor measurements which can be first transformed (outside the program) to cone or dihedral angle data can be processed. A cone angle is an angle between the spin axis and a known direction line in space; a dihedral angle is an angle between two planes formed by the spin axis and each of two known direction lines. Many different kinds of sensor data can be transformed to these angles, which in turn constitute the actual program inputs, so that the subroutine can be applied without change to a variety of satellite missions. Either a constant or dynamic spin axis model can be handled. The program is also capable of solving for fixed biases in the input angles, in addition to the spin axis attitude solution.

Werking, R. D.↗
NASA NTRS1975

A random search algorithm for laboratory computers

The small laboratory computer is ideal for experimental control and data acquisition. Postexperimental data processing is often performed on large computers because of the availability of sophisticated programs, but costs and data compatibility are negative factors. Parameter optimization can be accomplished on the small computer, offering ease of programming, data compatibility, and low cost. A previously proposed random-search algorithm ('random creep') was found to be very slow in convergence. A method is proposed (the 'random leap' algorithm) which starts in a global search mode and automatically adjusts step size to speed convergence. A FORTRAN executive program for the random-leap algorithm is presented which calls a user-supplied function subroutine. An example of a function subroutine is given which calculates maximum-likelihood estimates of receiver operating-characteristic parameters from binary response data. Other applications in parameter estimation, generalized least squares, and matrix inversion are discussed.

Curry, R. E.↗
NASA NTRS1977

Improved ground truth geoid for the GEOS-3 calibration area

The purpose of this investigation is to develop methods and procedures are reported for computing a detailed geoid to be used as geodetic ground truth for the calibration and verification of GEOS-3 altimeter data. The technique developed is based on rectifying the best available detailed geoid so that the rectified geoid will have correct scale, orientation, shape and position with respect to the geocenter. The approach involved the development of a mathematical model based on a second degree polynomial, in rectangular Cartesian coordinates, describing the geoid undulations at the control stations. A generalized least squares solution was obtained for the polynomial which describes the variation of the undulation differences between the control stations geoid and the gravimetric geoid. Three rectified geoid were determined. These geoids correspond to three sets of tracking station data: (1) WFC/C-band data; (2) GSFC/C-band data; and (3) OSU-275 data. The absolute accuracy of these rectified geoids is linearly correlated with the uncertainties of the tracking station coordinates and, to a certain extent, with those of the detailed geoid being rectified.

Mourad, A. G.↗
NASA NTRS1979

Error in total ozone measurements arising from aerosol attenuation

A generalized least squares method for deducing both total ozone and aerosol extinction spectrum parameters from Dobson spectrophotometer measurements was developed. An error analysis applied to this system indicates that there is little advantage to additional measurements once a sufficient number of line pairs have been employed to solve for the selected detail in the attenuation model. It is shown that when there is a predominance of small particles (less than about 0.35 microns in diameter) the total ozone from the standard AD system is too high by about one percent. When larger particles are present the derived total ozone may be an overestimate or an underestimate but serious errors occur only for narrow polydispersions.

Thomas, R. W. L.↗
NASA NTRS1986

A recursive solution for a fading memory filter derived from Kalman filter theory

A simple recursive solution for a class of fading memory tracking filters is presented. A fading memory filter provides estimates of filter states based on past measurements, similar to a traditional Kalman filter. Unlike a Kalman filter, an exponentially decaying weight is applied to older measurements, discounting their effect on present state estimates. It is shown that Kalman filters and fading memory filters are closely related solutions to a general least squares estimator problem. Closed form filter transfer functions are derived for a time invariant, steady state, fading memory filter. These can be applied in loop filter implementation of the Deep Space Network (DSN) Advanced Receiver carrier phase locked loop (PLL).

Statman, J. I.↗
NASA NTRS2002

Experimental Investigations of Non-Stationary Properties In Radiometer Receivers Using Measurements of Multiple Calibration References

Radiometers must be periodically calibrated because the receiver response fluctuates. Many techniques exist to correct for the time varying response of a radiometer receiver. An analytical technique has been developed that uses generalized least squares regression (LSR) to predict the performance of a wide variety of calibration algorithms. The total measurement uncertainty including the uncertainty of the calibration can be computed using LSR. The uncertainties of the calibration samples used in the regression are based upon treating the receiver fluctuations as non-stationary processes. Signals originating from the different sources of emission are treated as simultaneously existing random processes. Thus, the radiometer output is a series of samples obtained from these random processes. The samples are treated as random variables but because the underlying processes are non-stationary the statistics of the samples are treated as non-stationary. The statistics of the calibration samples depend upon the time for which the samples are to be applied. The statistics of the random variables are equated to the mean statistics of the non-stationary processes over the interval defined by the time of calibration sample and when it is applied. This analysis opens the opportunity for experimental investigation into the underlying properties of receiver non stationarity through the use of multiple calibration references. In this presentation we will discuss the application of LSR to the analysis of various calibration algorithms, requirements for experimental verification of the theory, and preliminary results from analyzing experiment measurements.

Racette, Paul↗
NASA NTRS1986

Study of a homotopy continuation method for early orbit determination with the Tracking and Data Relay Satellite System (TDRSS)

A recent mathematical technique for solving systems of equations is applied in a very general way to the orbit determination problem. The study of this technique, the homotopy continuation method, was motivated by the possible need to perform early orbit determination with the Tracking and Data Relay Satellite System (TDRSS), using range and Doppler tracking alone. Basically, a set of six tracking observations is continuously transformed from a set with known solution to the given set of observations with unknown solutions, and the corresponding orbit state vector is followed from the a priori estimate to the solutions. A numerical algorithm for following the state vector is developed and described in detail. Numerical examples using both real and simulated TDRSS tracking are given. A prototype early orbit determination algorithm for possible use in TDRSS orbit operations was extensively tested, and the results are described. Preliminary studies of two extensions of the method are discussed: generalization to a least-squares formulation and generalization to an exhaustive global method.

Smith, R. L.↗
NASA NTRS1979

Satellite interferometer as an advanced navigation/communication system

An orthogonal system of interferometry baselines aboard a geostationary satellite is proposed to be used for navigation, surveillance, and traffic control. Position determination is based on the relationship between the measured phase differences, the known and unknown transmitter locations, and other systematic error model parameters using a generalized least-squares estimation procedure. It is shown that position accuracy is critically dependent on the baseline length and on the magnitude of the random component of the measuring errors. Several trade-offs with respect to the system accuracies, the number of users, data rates, hardware costs, etc., are examined. Simultaneous use of the same channels for navigation and data transfer is found feasible and five candidate or 'strawman' interferometer system proposals are presented. It is concluded that the system's performance is excellent in comparison with present or planned systems, and that it is viable in meeting potential user's requirements.

Gopalapillai, G. S.↗
NASA NTRS1989

Automatic voice recognition using traditional and artificial neural network approaches

The main objective of this research is to develop an algorithm for isolated-word recognition. This research is focused on digital signal analysis rather than linguistic analysis of speech. Features extraction is carried out by applying a Linear Predictive Coding (LPC) algorithm with order of 10. Continuous-word and speaker independent recognition will be considered in future study after accomplishing this isolated word research. To examine the similarity between the reference and the training sets, two approaches are explored. The first is implementing traditional pattern recognition techniques where a dynamic time warping algorithm is applied to align the two sets and calculate the probability of matching by measuring the Euclidean distance between the two sets. The second is implementing a backpropagation artificial neural net model with three layers as the pattern classifier. The adaptation rule implemented in this network is the generalized least mean square (LMS) rule. The first approach has been accomplished. A vocabulary of 50 words was selected and tested. The accuracy of the algorithm was found to be around 85 percent. The second approach is in progress at the present time.

Botros, Nazeih M.↗
NASA NTRS1989

Multiple boundary condition testing error analysis

Techniques for interpreting data from multiple-boundary-condition (MBC) ground tests of large space structures are developed analytically and demonstrated. The use of MBC testing to validate structures too large to stand alone on the ground is explained; the generalized least-squares mass and stiffness curve-fitting methods typically applied to MBC test data are reviewed; and a detailed error analysis is performed. Consideration is given to sensitivity coefficients, covariance-matrix theory, the correspondence between test and analysis modes, constraints and step sizes, convergence criteria, and factor-analysis theory. Numerical results for a simple beam problem are presented in tables and briefly characterized. The improved error-updating capabilities of MBC testing are confirmed, and it is concluded that reasonably accurate results can be obtained using a diagonal covariance matrix.

Glaser, R. J.↗