Search NASA⌕ Search

SEARCH · Search NASA

Results for “Maximum likelihood estimation”

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 127 records · Page 7

Software models: A Bayesian approach to parameter estimation in the Jelenski-Moranda software reliability model

Maximum likelihood estimation procedures for the Jelinski-Moranda software reliability model often give misleading answers. A reparameterization and a Bayesian analysis eliminate some of the problems incurred by MLE methods and often give better predictions on sets of real and simulated data. Practical difficulties in estimating the initial number of errors N and the failure rate of each error phi by the method of maximum likelihood are: N, the MLE of N, is occasionally infinite (i.e., the routines for calculating N and phi do not converge). It is shown that N is finite sub i only if the regression line of the interevent times t sub i vs. i has positive slope. A serious problem is that often N approximates n, the sample size, and sometimes N = n. Thus the MLE predicts that the program is perfect even when it is far from being so. Only when almost all failures have been removed can N and phi be trusted near the end of debugging.

Littlewood, B.↗

System identification computational considerations.

System identification methods may be accurately characterized as inverse computational problems. Three approaches to the computational problem are considered. Equation error methods are based on regression analysis techniques. Attention is given to the maximum conditional likelihood estimate and the maximum unconditional likelihood (Bayesian) estimate. Gradient dependent algorithms for solving the minimization problems are discussed, taking into account Gaussian methods, quadratic model methods, and the homogeneous function model method of Jacobson and Oksman. Some advanced methods are also examined.

Bowles, R. L.↗

Digital adaptive control laws for VTOL aircraft

Honeywell has designed a digital self-adaptive flight control system for flight test in the VALT Research Aircraft (a modified CH-47). The final design resulted from a comparison of two different adaptive concepts: one based on explicit parameter estimates from a real-time maximum likelihood estimation algorithm and the other based on an implicit model reference adaptive system. The two designs are compared on the basis of performance and complexity.

Hartmann, G. L.↗

Estimating crop proportions from remotely sensed data

The classification/pixel-count method for estimating the proportion of wheat in each segment is theoretically biased even if all distributional assumptions are met. Alternative ways to estimate crop proportions are examined and their performance testing is considered. Topics covered include general linear functional estimates, the method of moments, and maximum likelihood estimators.

Feiveson, A. H.↗

Combining Earth System Modeling and Machine Learning to Investigate Volcanic Sulfate Deposition in Polar Ice Cores

Volcanic eruptions emit large amounts of sulfur dioxide (SO2), water, and other chemicals into the atmosphere, both in the troposphere and the stratosphere. Most of the SO2 is converted to sulfate aerosol, which is eventually deposited following long-range transport. The deposits from large eruptions are potentially detectable in ice cores, but there are many cases in which sulfate layers have not been linked to their source volcanoes. As volcanoes can act as significant shocks to the global climate system, we are interested in locating these eruptions in order to increase understanding of the volcanic record. To narrow down the search, we performed 140 simulations of volcanic eruptions using the GISS ModelE Earth system model. We varied the latitude, longitude, Julian day, plume top, plume bottom, and injected SO2 and H2O amounts using a Latin hypercube sampling approach, and analyzed correlations between these parameters and sulfate depositions at ice core sites in Antarctica and Greenland. Using machine learning and parameter estimation, we generated probability distributions and maximum likelihood estimates for the parameters given sulfate deposition data, which can predict latitude with some skill. We find that the volcano latitude and SO2 content are best correlated with sulfate depositions at each pole, while longitude, Julian day, and H2O have small or insignificant effects. Plume altitude and thickness are important because they determine how much of the SO2 is injected into the stratosphere, which has implications for sulfur transport and lifetimes.

Earth system models↗

Structure-aware Initialization via Numerical Continuation and Informed Priors

Scientific machine learning (SciML) often operates in ill-conditioned, weakly identifiable regimes due to limited data or indirect observations. In such settings, optimization and inference are highly sensitive to the starting point, making initialization--often under-reported--a consequential degree of freedom. Random initialization is not a neutral default as it induces an implicit prior over candidate solutions and can systematically bias the result, producing large run-to-run variability. Here, we formalize this view by treating initialization as a hidden confounder in SciML and develop a unifying theory for structure-aware initialization via numerical continuation, constructing warm starts from related problem instances. Across representative tasks, including physics-informed neural networks, maximum likelihood estimation, and variational inference, warm starts have been shown to consistently reduce optimization effort and improve reliability.

Data integrity↗

Approximate estimation for systems with quantized data.

Estimation of the state of a nonlinear discrete-time system using quantized data is considered. An exact solution for the maximum likelihood estimate is expressed as the solution of a nonlinear two-point boundary-value problem. Approximate recursive solutions for both the maximum likelihood and the conditional-mean estimates are obtained. The results of Monte-Carlo simulations are presented in which the performance of these two algorithms is compared with that of a Kalman filter in which the quantization error is approximated by white noise.-

Clements, K. A.↗

Digital adaptive controllers for VTOL vehicles. Volume 1: Concept evaluation

A digital self-adaptive flight control system was developed for flight test in the VTOL approach and landing technology (VALT) research aircraft (a modified CH-47 helicopter). The control laws accept commands from an automatic on-board guidance system. The primary objective of the control laws is to provide good command-following with a minimum cross-axis response. Three attitudes and vertical velocity are separately commanded. Adaptation of the control laws is based on information from rate and attitude gyros and a vertical velocity measurement. The final design resulted from a comparison of two different adaptive concepts--one based on explicit parameter estimates from a real-time maximum-likelihood estimation algorithm, the other based on an implicit model reference adaptive system. The two designs were compared on the basis of performance and complexity.

Hartmann, G. L.↗

Using known map category marginal frequencies to improve estimates of thematic map accuracy

By means of two simple sampling plans suggested in the accuracy-assessment literature, it is shown how one can use knowledge of map-category relative sizes to improve estimates of various probabilities. The fact that maximum likelihood estimates of cell probabilities for the simple random sampling and map category-stratified sampling were identical has permitted a unified treatment of the contingency-table analysis. A rigorous analysis of the effect of sampling independently within map categories is made possible by results for the stratified case. It is noted that such matters as optimal sample size selection for the achievement of a desired level of precision in various estimators are irrelevant, since the estimators derived are valid irrespective of how sample sizes are chosen.

Card, D. H.↗

Selected stability and control derivatives from the first three Space Shuttle entries

Stability and control derivative estimates obtained from the first three Space Shuttle entries are presented. The derivative estimates were obtained using the established modified maximum likelihood estimation program (MMLE3). The method of analysis used by the MMLE3 program is reviewed, the Shuttle configuration and data system are described, and the various types of maneuvers analyzed from the three entries are illustrated. Finally, the paper presents selected derivative results and compares them with predictions. Most of the flight-derived estimates agreed fairly well with predictions, considering the lack of experience in these new flight regimes. The most notable exception was the aerodynamic interference caused by firing the reaction control jets in the atmosphere. The flight results showed this interference to be considerably smaller than predicted.

Maine, R. E.↗

Timoshenko beam modeling for parameter estimation of NASA Mini-Mast truss

In this paper a distributed parameter model for the estimation of modal characteristics of NASA Mini-Mast truss is proposed. A closed-form solution of the Timoshenko beam equation, for a uniform cantilevered beam with two concentrated masses, is derived so that the procedure and the computational effort for the estimation of modal characteristics are improved. A maximum likelihood estimator for the Timoshenko beam model is also developed. The resulting estimates from test data by using Timoshenko beam model are found to be comparable to those derived from other approaches.

Shen, Ji Y.↗

Estimating residual fault hitting rates by recapture sampling

For the recapture debugging design introduced by Nayak (1988) the problem of estimating the hitting rates of the faults remaining in the system is considered. In the context of a conditional likelihood, moment estimators are derived and are shown to be asymptotically normal and fully efficient. Fixed sample properties of the moment estimators are compared, through simulation, with those of the conditional maximum likelihood estimators. Properties of the conditional model are investigated such as the asymptotic distribution of linear functions of the fault hitting frequencies and a representation of the full data vector in terms of a sequence of independent random vectors. It is assumed that the residual hitting rates follow a log linear rate model and that the testing process is truncated when the gaps between the detection of new errors exceed a fixed amount of time.

Lee, Larry↗

Accuracy of Aerodynamic Model Parameters Estimated from Flight Test Data

An important put of building mathematical models based on measured date is calculating the accuracy associated with statistical estimates of the model parameters. Indeed, without some idea of this accuracy, the parameter estimates themselves have limited value. An expression is developed for computing quantitatively correct parameter accuracy measures for maximum likelihood parameter estimates when the output residuals are colored. This result is important because experience in analyzing flight test data reveals that the output residuals from maximum likelihood estimation are almost always colored. The calculations involved can be appended to conventional maximum likelihood estimation algorithms. Monte Carlo simulation runs were used to show that parameter accuracy measures from the new technique accurately reflect the quality of the parameter estimates from maximum likelihood estimation without the need for correction factors or frequency domain analysis of the output residuals. The technique was applied to flight test data from repeated maneuvers flown on the F-18 High Alpha Research Vehicle. As in the simulated cases, parameter accuracy measures from the new technique were in agreement with the scatter in the parameter estimates from repeated maneuvers, whereas conventional parameter accuracy measures were optimistic.

Morelli, Eugene A.↗

Determining the Accuracy of Aerodynamic Model Parameters Estimated from Flight Test Data

An important part of building mathematical models based on measured data is calculating the accuracy associated with statistical estimates of the model parameters. Indeed, without some idea of this accuracy, the parameter estimates themselves have limited value. In this work, an expression for computing quantitatively correct parameter accuracy measures for maximum likelihood parameter estimates with colored residuals is developed and validated. This result is important because experience in analyzing flight test data reveals that the output residuals from maximum likelihood estimation are almost always colored. The calculations involved can be appended to conventional maximum likelihood estimation algorithms. Monte Carlo simulation runs were used to show that parameter accuracy measures from the new technique accurately reflect the quality of the parameter estimates from maximum likelihood estimation without the need for correction factors or frequency domain analysis of the output residuals. The technique was applied to flight test data from repeated maneuvers flown on the F-18 High Alpha Research Vehicle (HARV). As in the simulated cases, parameter accuracy measures from the new technique were in agreement with the scatter in the parameter estimates from repeated maneuvers, while conventional parameter accuracy measures were optimistic.

Morelli, Eugene A.↗

A new criterion for modeling systems

The criterion that is proposed is an expected value of the mean square response error as an alternative to testing a model against new data. Modeling with respect to this new criterion does not change the estimate for a given model format from a maximum likelihood estimate or mean square response error estimate. The new criterion does, however, provide a means of comparing models with different formats and varying complexity. A numerical example is used to illustrate the application of the proposed criteria and the problem of searching for the best model. For all but the most trivial system identification problems, it is shown that a prohibitive number of combinations of terms of the model must be investigated to ensure the final model is best.

Taylor, L. W., Jr.↗

Estimation of the differential pulse propagation times in two-color laser ranging systems

Two-color laser ranging systems can be used to determine the atmospheric delay by measuring the difference in propagation times between two optical pulses transmitted at two different wavelengths. In this paper, the performance of a cross-correlation technique for estimating the differential propagation time is analyzed by considering both speckle and shot noise. For the flat diffuse targets, the differential timing accuracy is highly dependent on the receiver bandwidth and the characteristics of the time-resolved speckle. At low signal levels, when the receiver bandwidth is chosen properly, the performance of the correlation estimator is comparable with that of the maximum-likelihood estimator. At high signal levels, however, speckle places a fundamental limit on the performance of the correlation estimator. For the cube-corner reflector arrays, timing performance is dominated by partially developed speckle. The differential propagation time cannot by resolved to better than the pulse widths of the received signals.

Im, K. E.↗

Transonic stability and control of aircraft using CFD methods

Implementation of a capability to calculate longitudinal short-period response in the CAP-TSD (Computational Aeroelasticity Program - Transonic Small Disturbance) finite-difference code is described. The code, developed recently at the NASA Langley Research Center, is capable of solving steady and unsteady flows about complete aircraft configurations and is used primarily for aeroelastic calculations in the critical transonic speed range. The longitudinal short-period equations of motion in state-space form have been coupled to the time-accurate lift and moment calculated by the program. Transient responses to an elevator pulse for free-flying aircraft demonstrate the new capability. A trim routine is also added to the code to obtain trim automatically during steady-state flow field convergence. Stability and control derivatives are estimated from the calculated transient response by a maximum likelihood estimation program. Results for a fighter configuration and a general aviation configuration are presented to assess the capability.

Vinh, Lam-Son↗

X-Ray Detection and Processing Models for Spacecraft Navigation and Timing

The current primary method of deepspace navigation is the NASA Deep Space Network (DSN). High-performance navigation is achieved using Delta Differential One-Way Range techniques that utilize simultaneous observations from multiple DSN sites, and incorporate observations of quasars near the line-of-sight to a spacecraft in order to improve the range and angle measurement accuracies. Over the past four decades, x-ray astronomers have identified a number of xray pulsars with pulsed emissions having stabilities comparable to atomic clocks. The x-ray pulsar-based navigation and time determination (XNAV) system uses phase measurements from these sources to establish autonomously the position of the detector, and thus the spacecraft, relative to a known reference frame, much as the Global Positioning System (GPS) uses phase measurements from radio signals from several satellites to establish the position of the user relative to an Earth-centered fixed frame of reference. While a GPS receiver uses an antenna to detect the radio signals, XNAV uses a detector array to capture the individual xray photons from the x-ray pulsars. The navigation solution relies on detailed xray source models, signal processing, navigation and timing algorithms, and analytical tools that form the basis of an autonomous XNAV system. Through previous XNAV development efforts, some techniques have been established to utilize a pulsar pulse time-of-arrival (TOA) measurement to correct a position estimate. One well-studied approach, based upon Kalman filter methods, optimally adjusts a dynamic orbit propagation solution based upon the offset in measured and predicted pulse TOA. In this delta position estimator scheme, previously estimated values of spacecraft position and velocity are utilized from an onboard orbit propagator. Using these estimated values, the detected arrival times at the spacecraft of pulses from a pulsar are compared to the predicted arrival times defined by the pulsar s pulse timing model. A discrepancy provides an estimate of the spacecraft position offset, since an error in position will relate to the measured time offset of a pulse along the line of sight to the pulsar. XNAV researchers have been developing additional enhanced approaches to process the photon TOAs to arrive at an estimate of spacecraft position, including those using maximum-likelihood estimation, digital phase locked loops, and "single photon processing" schemes that utilize all available time data associated with each photon. Using pulsars from separate, non-coplanar locations provides range and range-rate measurements in each pulsar s direction. Combining these different pulsar measurements solves for offsets in position and velocity in three dimensions, and provides accurate overall navigation for deep space vehicles.

Sheikh, Suneel↗