Search NASA⌕ Search

SEARCH · Search NASA

Results for “Linearization 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 1,585 records · Page 88

The Impact of Soil Reflectance on the Quantification of the Green Vegetation Fraction from NDVI

The green vegetation fraction (Fg) is an important climate and hydrologic model parameter. A common method to calculate Fg is to create a simple linear mixing rnodeP between two NDVI endmembers: bare soil NDVI (NDVI(sub o)) and full vegetation NDVI (NDVI(sub infinity)). Usually it is assumed that NDVI(sub o), is close to zero (NDVI(sub o) approx.-0.05) and is generally chosen from the lowest observed NDVI values. However, the mean soil NDVI computed from 2906 samples is much larger (NDVI=0.2) and is highly variable (standard deviation=O. 1). We show that the underestimation of NDVI(sub o) yields overestimations of Fg. The largest errors occur in grassland and shrubland areas. Using parameters for NDVI(sub o) and NDVI(sub infinity) derived from global scenes yields overestimations of Fg ((Delta) Fg*) that are larger than 0.2 for the majority of U.S. land cover types when pixel NDVI values are 0.2<NDVI(sub pixel)<0.4. Figure 1 shows how the Fg overestimation varies for the most common land cover types in the conterminous U.S. for typical seasonal NDVI values. When using conterminous U.S. scenes to derive NDV(sub o) and NDVI(sub infinity), the overestimation is less (0.10-0.17 for 0.2<NDVI(sub pixel)<0.4). As a result, parts of the conterminous U.S. are affected at different times of the year depending on the local seasonal NDVI cycle. We propose using global databases of NDVI(sub o) along with information on historical NDVI(sub pixel) values to compute a statistically most-likely estimate of Fg (Fg*). Using in situ measurements made at the Sevilleta LTER, we show that this approach yields better estimates of Fg than using global invariant NDVI(sub o) values estimated from whole scenes (Figure 2). At the two studied sites, the Fg estimate was adjusted by 52% at the grassland and 86% at the shrubland. More significant advances will require information on spatial distribution of soil reflectance.

Montandon, L. M.↗

Developing a Fundamental Model for an Integrated GPS/INS State Estimation System with Kalman Filtering

This work will demonstrate the integration of sensor and system dynamic data and their appropriate models using an optimal filter to create a robust, adaptable, easily reconfigurable state (motion) estimation system. This state estimation system will clearly show the application of fundamental modeling and filtering techniques. These techniques are presented at a general, first principles level, that can easily be adapted to specific applications. An example of such an application is demonstrated through the development of an integrated GPS/INS navigation system. This system acquires both global position data and inertial body data, to provide optimal estimates of current position and attitude states. The optimal states are estimated using a Kalman filter. The state estimation system will include appropriate error models for the measurement hardware. The results of this work will lead to the development of a "black-box" state estimation system that supplies current motion information (position and attitude states) that can be used to carry out guidance and control strategies. This black-box state estimation system is developed independent of the vehicle dynamics and therefore is directly applicable to a variety of vehicles. Issues in system modeling and application of Kalman filtering techniques are investigated and presented. These issues include linearized models of equations of state, models of the measurement sensors, and appropriate application and parameter setting (tuning) of the Kalman filter. The general model and subsequent algorithm is developed in Matlab for numerical testing. The results of this system are demonstrated through application to data from the X-33 Michael's 9A8 mission and are presented in plots and simple animations.

Canfield, Stephen↗

Optical properties of intact leaves for estimating chlorophyll concentration

Changes in leaf chlorophyll content can serve as relative indicators of plant vigor and environmental quality. This study identified reflectance, transmittance, and absorptance wavebands and band ratios within the 400- to 850-nm range for intact leaves that could be used to estimate extracted leaf chlorophyll per unit leaf area (areal concentration) with minimal error. Leaf optical properties along with chlorophyll a, b, and a + b concentrations were measured for the planar-leaved sweetgum (Liquidambar styraciflua L.), red maple (Acer rubrum L.), wild grape (Vitis rotundifolia Michx.), and switchcane [Arundinaria gigantea (Walter) Muhl.], and for needles of longleaf pine (Pinus palustris Miller). Generally, reflectance, transmittance, and absorptance corresponded most precisely with chlorophyll concentrations at wavelengths near 700 nm, although regressions were also strong in the 550- to 625-nm range. A power function was superior to a simple linear function in yielding low standard deviations of the estimate (s). When data were combined among the planar-leaved species, s values were low at approximately 50 mumol/m2 out of a 940 mumol/m2 range in chlorophyll a + b at best-fit wavelengths of 707 to 709 nm. Minimal s values for chlorophyll a + b ranged from 32 to 62 mumol/m2 across species when band ratios having numerator wavelengths of 693 to 720 nm were used with the application of a power function. Optimal denominator wavelengths for the band ratios were 850 nm for reflectance and transmittance and 400 nm for absorptance. This information can be applied in designing field portable chlorophyll meters and in the landscape-scale remote sensing of plant responses to the environment.

Chlorophyll/analysis↗

Using Strong Solar Coronal Emission Lines as Coronal Flux Proxies

A comparison of Skylab results with observations of the strong EUV lines of Fe XVI at 335 A and 361 A from the Goddard Solar EUV Rocket Telescope and Spectrograph (SERTS) flight of 1989 suggests that these lines, and perhaps others observed with SERTS, might offer good proxies for estimating the total coronal flux over important wavelength ranges. In this paper, we compare SERTS observations from a later, 1993 flight with simultaneous cospatial Yohkoh soft X-ray observations to test this suggestion over the energy range of the Soft X-ray Telescope (SXT) on Yohkoh. Both polynomial and power-law fits are obtained, and errors are estimated, for the SERTS lines of Fe XVI 335 A and 361 A, Fe XV 284 A and 417 A, and Mg IX 368 A. It is found that the power-law fits best cover the full range of solar conditions from quiet Sun through active region, though not surprisingly the 'cooler' Mg IX 368 A line proves to be a poor proxy. The quadratic polynomial fits yield fair agreement over a large range for all but the Mg IX line, but the linear fits fail conspicuously when extrapolated into the quiet Sun regime. The implications of this work for the He 11 304 A line formation problem are briefly considered. The paper concludes with a discussion of the value of these iron lines observed with SERTS for estimating stellar coronal fluxes, as observed for example with the EUVE satellite.

Falconer, David A.↗

Neural Network and Response Surface Methodology for Rocket Engine Component Optimization

The goal of this work is to compare the performance of response surface methodology (RSM) and two types of neural networks (NN) to aid preliminary design of two rocket engine components. A data set of 45 training points and 20 test points obtained from a semi-empirical model based on three design variables is used for a shear coaxial injector element. Data for supersonic turbine design is based on six design variables, 76 training, data and 18 test data obtained from simplified aerodynamic analysis. Several RS and NN are first constructed using the training data. The test data are then employed to select the best RS or NN. Quadratic and cubic response surfaces. radial basis neural network (RBNN) and back-propagation neural network (BPNN) are compared. Two-layered RBNN are generated using two different training algorithms, namely solverbe and solverb. A two layered BPNN is generated with Tan-Sigmoid transfer function. Various issues related to the training of the neural networks are addressed including number of neurons, error goals, spread constants and the accuracy of different models in representing the design space. A search for the optimum design is carried out using a standard gradient-based optimization algorithm over the response surfaces represented by the polynomials and trained neural networks. Usually a cubic polynominal performs better than the quadratic polynomial but exceptions have been noticed. Among the NN choices, the RBNN designed using solverb yields more consistent performance for both engine components considered. The training of RBNN is easier as it requires linear regression. This coupled with the consistency in performance promise the possibility of it being used as an optimization strategy for engineering design problems.

Vaidyanathan, Rajkumar↗

An Explicit Linear Filtering Solution for the Optimization of Guidance Systems with Statistical Inputs

The determination of optimum filtering characteristics for guidance system design is generally a tedious process which cannot usually be carried out in general terms. In this report a simple explicit solution is given which is applicable to many different types of problems. It is shown to be applicable to problems which involve optimization of constant-coefficient guidance systems and time-varying homing type systems for several stationary and nonstationary inputs. The solution is also applicable to off-design performance, that is, the evaluation of system performance for inputs for which the system was not specifically optimized. The solution is given in generalized form in terms of the minimum theoretical error, the optimum transfer functions, and the optimum transient response. The effects of input signal, contaminating noise, and limitations on the response are included. From the results given, it is possible in an interception problem, for example, to rapidly assess the effects on minimum theoretical error of such factors as target noise and missile acceleration. It is also possible to answer important questions regarding the effect of type of target maneuver on optimum performance.

Stewart, Elwood C.↗

Modal methods in optimal control synthesis

Efficient algorithms for solving linear smoother-follower problems with quadratic criteria are presented. For time-invariant systems, the algorithm consists of one backward integration of a linear vector equation and one forward integration of another linear vector equation. Furthermore, the backward and forward Riccati matrices can be expressed in terms of the eigenvalues and eigenvectors of the Euler-Lagrange equations. Hence, the gains of the forward and backward Kalman-Bucy filters and of the optimal state-feedback regulator can be determined without integration of matrix Riccati equations. A computer program has been developed, based on this method of determining the gains, to synthesize the optimal time-invariant compensator in the presence of random disturbance inputs and random measurement errors. The program also computes the rms state and control variables of the optimal closed-loop system.

Bryson, A. E., Jr.↗

Evaluation of electrolytic tilt sensors for measuring model angle of attack in wind tunnel tests

The results of a laboratory evaluation of electrolytic tilt sensors as potential candidates for measuring model attitude or angle of attack in wind tunnel tests are presented. The performance of eight electrolytic tilt sensors was compared with that of typical servo accelerometers used for angle-of-attack measurements. The areas evaluated included linearity, hysteresis, repeatability, temperature characteristics, roll-on-pitch interaction, sensitivity to lead-wire resistance, step response time, and rectification. Among the sensors being evaluated, the Spectron model RG-37 electrolytic tilt sensors have the highest overall accuracy in terms of linearity, hysteresis, repeatability, temperature sensitivity, and roll sensitivity. A comparison of the sensors with the servo accelerometers revealed that the accuracy of the RG-37 sensors was on the average about one order of magnitude worse. Even though a comparison indicates that the cost of each tilt sensor is about one-third the cost of each servo accelerometer, the sensors are considered unsuitable for angle-of-attack measurements. However, the potential exists for other applications such as wind tunnel wall-attitude measurements where the errors resulting from roll interaction, vibration, and response time are less and sensor temperature can be controlled.

Wong, Douglas T.↗

Numerical simulation of controlled large space structures

Large Space Structures do not have much damping, which necessitates the installation of a controller onto the structure. If the controller is improperly designed, the structure may become unstable and be destroyed. Since Large Space Structures are extremely expensive pieces of hardware, new controllers must not be tested first on the structure. They must first be tested in computer simulations. Until now, the usual procedure for simulating controlled Large Space Structures is to compute a reduced order modal representation of the structure and then apply the controller. However, this procedure entails modal truncation error. A new software package which is free from this error is currently under development within the Center for Space Construction. The more accurate finite element representation of the structure is used in the simulation, instead of the less accurate reduced order modal representation. This software also features an efficient matrix storage scheme, which effectively deals with the asymmetric system matrices which occur when control is added to the structure. Also, an integration algorithm was chosen so that the simulation is a reliable indicator of system stability or instability. The software package is fairly general in nature. Linearity of the finite element model and of the controller is the only assumption made. Actuator dynamics, sensor dynamics, noise, and disturbances can be handled by the package. In addition, output feedback of displacement, velocity, and/or acceleration signals can be simulated. Kalman state estimation was also implemented. This software was tested on a finite element model of a real Large Space Structure: The Mini-Mast Truss. Mini-Mast is a testbed at NASA-Langley which is currently under development. A 714 degree of freedom finite element model was computed, and a 19 state controller was designed for it. Torque wheel dynamics were added to the model, and the entire closed loop system was simulated with the software package.

Quan, Ralph↗

On the Possibility of Ill-Conditioned Covariance Matrices in the First-Order Two-Step Estimator

The first-order two-step nonlinear estimator, when applied to a problem of orbital navigation, is found to occasionally produce first step covariance matrices with very low eigenvalues at certain trajectory points. This anomaly is the result of the linear approximation to the first step covariance propagation. The study of this anomaly begins with expressing the propagation of the first and second step covariance matrices in terms of a single matrix. This matrix is shown to have a rank equal to the difference between the number of first step states and the number of second step states. Furthermore, under some simplifying assumptions, it is found that the basis of the column space of this matrix remains fixed once the filter has removed the large initial state error. A test matrix containing the basis of this column space and the partial derivative matrix relating first and second step states is derived. This square test matrix, which has dimensions equal to the number of first step states, numerically drops rank at the same locations that the first step covariance does. It is formulated in terms of a set of constant vectors (the basis) and a matrix which can be computed from a reference trajectory (the partial derivative matrix). A simple example problem involving dynamics which are described by two states and a range measurement illustrate the cause of this anomaly and the application of the aforementioned numerical test in more detail.

Garrison, James L.↗

On the validity of the double integrator approximation in deep space formation flying

Free-flying models are commonly used f o r path planning and open loop control design (i. e., guidance design) and translational feedback control design (i. e., control design) for deep space precision formation flying. The free flying model, essentially a double integrator, results from discarding small terms in the relative spacecraft equations of motion. While the magnitude of these discarded terms may be small, one must show that their dynamic effects are small as compared to the precision performance requirements. We do so by deriving a theoretical method for bounding the difference between the solution of a nonlinear truth model of the relative translational spacecraft dynamics and a Simplified linear time-invariant model. Presently, the method incorporates feedforward and static output feedback control. The method is applied to a Terrestrial Planet Finder- based example. Using only feedforward control (guidance) the free-flying model and a Hill- Clohessy- Wiltshire Equations-based model are shown to be accurate to 1 c m for up to 4 and 30 hours, respectively. Also shown is that the simplest free-flying model may not be sufficient for low-gain feedback control design-closed-loop tracking errors can be as large as 8 meters.

formation flying free flying model double integrat↗

Intrastep, Stage-Value Predictors for Diagonally-Implicit Runge-Kutta Methods

To better identify the necessary attributes of good stage-value predictors (SVPs), numerous SVPs are designed for an existing: ESDIRK4(3)7L[2]SA [26] and a new: ESDIRK4(3)8L[2]SA scheme.1 Both are stifflyaccurate, stage-order two, explicit, singly-diagonally implicit Runge–Kutta (ESDIRK) schemes. Tradeoffs are studied in the parameter spaces enforcing the constraints on accuracy, linear stability, nonlinear stability and coefficient size to determine which objectives correlate with effective predictors. The SVPs are tested on three challenging external aerodynamics problems [107 − 108 degrees of freedom (DoFs)], each with a different level of stiffness. The problems include two 3D airfoils simulations and one canonical turbulence simulation. All simulations use the compressible Navier-Stokes equations (CNSE). An entropy stable spectral collocation formulation is used for discretizing the spatial terms in the equations. Simulations are performed at a wide variety of temporal error tolerances. Problems that are sufficiently stiff (e.g., lax temporal error tolerances) benefit from SVPs designed with second-order accuracy and stability properties: A-stability, and L-stability, rather than high accuracy constraints. Simulations with modest stiffness (e.g., strict error tolerances) are better suited for SVPs designed using high accuracy constraints. Designing SVPs with enhanced stability properties is tedious but worthwhile. Simulation times are reduced with optimal SVPs by as much as 100% on some stages, with combined stepwise improvements of between 50 − 100% for both methods. A comparative study is performed with the two aforementioned methods as well as four other ESDIRKs. The newly designed ESDIRK4(3)8L[2]SA with γ ≈ 1/10, proves to be the most efficient of the six tested ESDIRK schemes simulating the CNSE.

Diagonally-Implicit Runge-Kutta↗

A triangular thin shell finite element: Linear analysis

The formulation of the linear stiffness matrix for a doubly-curved triangular thin shell element, using a modified potential energy principle, is described. The strain energy component of the potential energy is expressed in terms of displacements and displacement gradients by use of consistent Koiter strain-displacement equations. The element inplane and normal displacement fields are approximated by complete cubic polynomials. The interelement displacement admissibility conditions are met in the global representation by imposition of constraint conditions on the interelement boundaries; the constraints represent the modification of the potential energy. Errors due to the nonzero strains under rigid body motion are shown to be of small importance for practical grid refinements through performance of extensive comparison analyses.

Thomas, G. R.↗

An analytically linearized helicopter model with improved modeling accuracy

An analytically linearized model for helicopter flight response including rotor blade dynamics and dynamic inflow, that was recently developed, was studied with the objective of increasing the understanding, the ease of use, and the accuracy of the model. The mathematical model is described along with a description of the UH-60A Black Hawk helicopter and flight test used to validate the model. To aid in utilization of the model for sensitivity analysis, a new, faster, and more efficient implementation of the model was developed. It is shown that several errors in the mathematical modeling of the system caused a reduction in accuracy. These errors in rotor force resolution, trim force and moment calculation, and rotor inertia terms were corrected along with improvements to the programming style and documentation. Use of a trim input file to drive the model is examined. Trim file errors in blade twist, control input phase angle, coning and lag angles, main and tail rotor pitch, and uniform induced velocity, were corrected. Finally, through direct comparison of the original and corrected model responses to flight test data, the effect of the corrections on overall model output is shown.

Jensen, Patrick T.↗

Modified optimal control pilot model for computer-aided design and analysis

This paper presents the theoretical development of a modified optimal control pilot model based upon the optimal control model (OCM) of the human operator developed by Kleinman, Baron, and Levison. This model is input compatible with the OCM and retains other key aspects of the OCM, such as a linear quadratic solution for the pilot gains with inclusion of control rate in the cost function, a Kalman estimator, and the ability to account for attention allocation and perception threshold effects. An algorithm designed for each implementation in current dynamic systems analysis and design software is presented. Example results based upon the analysis of a tracking task using three basic dynamic systems are compared with measured results and with similar analyses performed with the OCM and two previously proposed simplified optimal pilot models. The pilot frequency responses and error statistics obtained with this modified optimal control model are shown to compare more favorably to the measured experimental results than the other previously proposed simplified models evaluated.

Davidson, John B.↗

Improving the performance of soft decision Viterbi decoding in a non-Gaussian environment through non-linear quantization

The performance of Viterbi decoding in a non-Gaussian environment is investigated using a nonlinear quantization strategy. The channel model consists of a convolutionally encoded BPSK signal transmitted to a satellite where it is corrupted with additive white Gaussian noise and pulsed radio frequency interference (RFI). The resultant signal is then passed through a satellite nonlinearity and transmitted to a ground station where it is coherently detected. Interleaving is assumed in order to make the channel memoryless. The presence of RFI makes the channel statistics non-Gaussian, leading to a nonlinear log-likelihood function. A near optimum quantization scheme is found by maximizing a channel parameter, or by matching the quantizer to the log-likelihood function in a mean square error sense. Bit error rate performance improvement is achieved by using such nonlinear quantization.

Mcgregor, D. N.↗

Vector Sum Excited Linear Prediction (VSELP) speech coding at 4.8 kbps

Code Excited Linear Prediction (CELP) speech coders exhibit good performance at data rates as low as 4800 bps. The major drawback to CELP type coders is their larger computational requirements. The Vector Sum Excited Linear Prediction (VSELP) speech coder utilizes a codebook with a structure which allows for a very efficient search procedure. Other advantages of the VSELP codebook structure is discussed and a detailed description of a 4.8 kbps VSELP coder is given. This coder is an improved version of the VSELP algorithm, which finished first in the NSA's evaluation of the 4.8 kbps speech coders. The coder uses a subsample resolution single tap long term predictor, a single VSELP excitation codebook, a novel gain quantizer which is robust to channel errors, and a new adaptive pre/postfilter arrangement.

Gerson, Ira A.↗

A statistical model of diurnal variation in human growth hormone

The diurnal pattern of growth hormone (GH) serum levels depends on the frequency and amplitude of GH secretory events, the kinetics of GH infusion into and clearance from the circulation, and the feedback of GH on its secretion. We present a two-dimensional linear differential equation model based on these physiological principles to describe GH diurnal patterns. The model characterizes the onset times of the secretory events, the secretory event amplitudes, as well as the infusion, clearance, and feedback half-lives of GH. We illustrate the model by using maximum likelihood methods to fit it to GH measurements collected in 12 normal, healthy women during 8 h of scheduled sleep and a 16-h circadian constant-routine protocol. We assess the importance of the model components by using parameter standard error estimates and Akaike's Information Criterion. During sleep, both the median infusion and clearance half-life estimates were 13.8 min, and the median number of secretory events was 2. During the constant routine, the median infusion half-life estimate was 12.6 min, the median clearance half-life estimate was 11.7 min, and the median number of secretory events was 5. The infusion and clearance half-life estimates and the number of secretory events are consistent with current published reports. Our model gave an excellent fit to each GH data series. Our analysis paradigm suggests an approach to decomposing GH diurnal patterns that can be used to characterize the physiological properties of this hormone under normal and pathological conditions.

Non-NASA Center↗