Initial Parameter Estimates Research Articles

A combined neural network/gradient‐based approach for the identification of constitutive model parameters using self‐boring pressuremeter tests

AbstractThis paper presents a numerical procedure of material parameter identification for the coupled hydro‐mechanical boundary value problem (BVP) of the self‐boring pressuremeter test (SBPT) in clay. First, the neural network (NN) technique is applied to obtain an initial estimate of model parameters, taking into account the possible drainage conditions during the expansion test. This technique is used to avoid potential pitfalls related to the conventional gradient‐based optimization techniques, considered here as a corrector that improves predicted parameters. Parameter identification based on measurements obtained through the pressuremeter expansion test and two types of holding tests is illustrated on the Modified Cam clay model. NNs are trained using a set of test samples, which are generated by means of finite element simulations of SBPT. The measurements obtained through expansion and consolidation tests are normalized so that NN predictors operate independently of the testing depth. Examples of parameter determination are demonstrated on both numerical and field data. The efficiency of the combined parameter identification in terms of accuracy, effectiveness and computational effort is also discussed. Copyright © 2008 John Wiley & Sons, Ltd.

Inclusion Identification by Inverse Application of Boundary Element Method, Genetic Algorithm and Conjugate Gradient Method

Characterization of the interior of an inhomogeneous body using displacement measurements obtained by conducting a simple tension test was investigated. A homogeneous elastic solid body is assumed to have another solid body with arbitrary shape hidden inside it. The shape and physical properties of this inclusion were unknown. The Boundary Element Method (BEM) coupled with a mete heuristic Genetic Algorithm (GA) and Conjugate Gradient Method (CGM) was used in this characterization problem. A fitness function, which is the summation of squared differences between the measured displacements and calculated displacements at the same locations on the boundary, is minimized using GA and CGM. GA is used to find a good initial estimate of the unknown parameters and the CGM was used as the hybrid function to get converged values of the unknown parameters. For the cases that the fitness function fluctuates severely, a regularization function was added to CGM in order to dampen the fluctuations.

Electron density and electron neutral collision frequency in the ionosphere using plasma impedance probe measurements

Swept Impedance Probe measurements in a sporadic E layer observed during the Sudden Atomic Layer (SAL) sounding rocket mission are analyzed to obtain absolute electron densities and electron neutral collision frequencies accurately. Three sets of upleg and downleg impedance data are selected for the analysis. Initial estimates of the plasma parameters are obtained through a least mean square fit of the measured impedance data against the analytical impedance formula ZB(f) of Balmain (1969). These initial parameters are used as a starting point to drive a finite difference computational model of an antenna immersed in a plasma called PF‐FDTD. The parameters are then tuned until a close fit is obtained between the measured impedance data and the numerical impedance data calculated by the PF‐FDTD simulation. The electron densities obtained from the simulation were close to those obtained from the IRI 2001 model. The electron neutral collision frequencies obtained from the more accurate PF‐FDTD simulation were up to 20% lower than the values predicted by Balmain's formula. The obtained collision frequencies are also lower than the quiet time values predicted by Schunk and Nagy (2000) when used in conjunction with neutral densities and electron temperature from the Mass Spectrometer Incoherent Scatter Radar Extended‐90 model.

Multi-Scale Adaptive Sampling with Mobile Agents for Mapping of Forest Fires

The use of robotics in distributed monitoring applications requires wireless sensors that are deployed efficiently. A very important aspect of sensor deployment includes positioning them for sampling at locations most likely to yield information about the spatio-temporal field of interest, for instance, the spread of a forest fire. In this paper, we use mobile robots (agents) that estimate the time-varying spread of wildfires using a distributed multi-scale adaptive sampling strategy. The proposed parametric sampling algorithm, "EKF-NN-GAS" is based on neural networks, the extended Kalman filter (EKF), and greedy heuristics. It combines measurements arriving at different times, taken at different scale lengths, such as from ground, airborne, and spaceborne observation platforms. One of the advantages of our algorithm is the ability to incorporate robot localization uncertainty in addition to sensor measurement and field parameter uncertainty into the same EKF model. We employ potential fields, generated naturally from the estimated fire field distribution, in order to generate fire-safe trajectories that could be used to rescue vehicles and personnel. The covariance of the EKF is used as a quantitative information measure for sampling locations most likely to yield optimal information about the sampled field distribution. Neural net training is used infrequently to generate initial low resolution estimates of the fire spread parameters. We present simulation and experimental results for reconstructing complex spatio-temporal forest fire fields "truth models", approximated by radial basis function (RBF) parameterizations. When compared to a conventional raster scan approach, our algorithm shows a significant reduction in the time necessary to map the fire field.

A multi-species, process based vegetation simulation module to simulate successional forest regrowth after forest disturbance in daily time step hydrological transport models

To simulate the effects of tree harvest on boreal forest catchment hydrology, the vegetation growth model within the soil and water assessment tool (SWAT) must reproduce the successional stages of forest reestablishment. The agricultural land management alternatives with numerical assessment criteria (ALMANAC), a multi-species growth model, was modified to simulate vegetation regeneration on forest sites after harvest (ALMANACBF). The model uses similar principles of vegetation growth as the current vegetation model in SWAT, and input requirements are consistent with typical forest inventory databases. This article describes the algorithms integrated into the ALMANACBF model to simulate successional stages in forest growth, provides initial estimates of parameters required to simulate multi-species forest succession, and presents examples of the type of variability in vegetation growth scenarios that these algorithms can reproduce. The model structure and modelling approach shows promise as a tool for foresters to evaluate how patterns and timing of forest management activities influence forest watershed hydrology.

A Real-Time 4-Stream MIMO-OFDM Transceiver: System Design, FPGA Implementation, and Characterization

When designing complex communication systems, such as MIMO-OFDM transceivers, prototypes have become an important tool for understanding the implementation trade-offs and the system behavior. This paper presents a real-time FPGA prototype for a 4-stream MIMO-OFDM transceiver capable of transmitting 216 Mbit/s in 20 MHz bandwidth. The paper covers all parts of the system from RF to channel decoding and considers both algorithm and implementation aspects. In particular, we discuss the initial parameter estimation, channel estimation, MIMO detection, parameter tracking, and channel decoding. FPGA implementation results are reported along with measurements that demonstrate the throughput of spatial multiplexing with four spatial streams.

Poster - Thurs Eve-28: New brain diffusion analysis method: White matter grey matter dissasociation.

Diffusion MR studies are often used to investigate the physical properties of brain tissues (1, 2). It is known that a full characterisation of the diffusion decay for brain could give valuable information about the structural organisation of cerebral tissue. The significance of the present diffusion decay study lies in the combination of three novel procedures to provide a better characterization of the diffusion decay: i) the acquisition of a large number of b-values (96 b-values up to 10,000 s/mm2 ), ii) the application of a noise correction technique (3) to the acquired data, and iii) the use of a Non Negative Least Squares (NNLS) fitting algorithm to evaluate the diffusion coefficients. The presence of noise in magnitude MR images can affect the calculation of the diffusion parameters (4) and therefore a noise correction technique (3) is applied. The NNLS algorithm is used to fit the corrected data instead of the more commonly used Levenberg-Marquardt algorithm since the NNLS algorithm does not require the number of components to be specified, nor does it need initial estimates of the fitting parameters as input; thus giving it more versatility as a fitting tool for the diffusion decay. The results indicate that the diffusion decays in grey and white matter have one and two components, respectively. Consequently, the short diffusion component in white matter (Fig. 1.c) can be used as a tool in the disassociation of white and grey matter tissues.

Increasing the usefulness of additive spline models by knot removal

Modern techniques for fitting generalized additive models mostly rely on basis expansions of covariates using a large number of basis functions and penalized estimation of parameters. For example, a mixed model approach is used to fit a model for children’s lung function that allows for non-linear influence of several covariates available in a substantial data set. While the resulting model is expected to have good prediction performance, its handling beyond simple visual presentation is problematic. It is shown how the number basis functions of the underlying B-spline representation can be reduced by knot removal techniques without refitting, while preserving the shape of the fitted functions. The condition for exact knot removal is extended towards approximate knot removal by incorporating the covariance matrix of the initial parameter estimates, resulting in considerable simplification of the model. Covariance matrices for the transformed parameter estimates are provided. It is demonstrated that enforcing the knot removal condition during estimation leads to the difference penalties employed in the P-spline approach for estimation of B-spline coefficients, and therefore provides a further justification for this type of penalty. A final transform to a truncated power basis provides a simple equation for the model. This increases transportability, while retaining properties of the initial fit such as good prediction performance.

Determining dynamic viscoelastic properties without time‐temperature shifting

A novel unbiased procedure to analyze dynamic mechanical data of rheologically simple viscoelastic polymers by modeling the data with the Havriliak and Negami (HN) equation is described. The real and imaginary parts of the HN equation are used to solve for the global frequency‐time dependent parameter (ωτ) for all the data thereby uniquely determining the time‐ temperature shift function. Displaying the experimental data in the form of a wicket or Argand diagram provides initial estimates for the HN parameters. An unbiased error analysis is performed to minimize the difference between experimental and calculated complex viscoelastic values. Finally the characteristic relaxation time, τb> , is determined by minimizing the error between the calculated data and experimental data at the reference temperature . Using this procedure, the complete master curve is generated without the need for overlapping frequency data and the procedural error and operator bias associated with time‐temperature shifti...

Symbolic Vector Dynamics Approach to Initial Condition and Control Parameters Estimation of Coupled Map Lattices

In this paper, we extend symbolic dynamics, a standard analytical method for 1-D chaotic map, and initiate a solution to the problem of estimation in coupled map lattices (CMLs) by introducing the symbolic vector dynamics. We develop a novel technique for estimating initial conditions. We also expand the applicable scope of a word-lifting technique for parameter estimation from a 1-D chaotic map to CMLs. Both theoretical and experimental results show that those algorithms can construct one-to-one correspondence between the set of global orbits and the set of admissible codes. Therefore, we provide novel analytical techniques for understanding turbulences in CMLs.

Individualized performance prediction of sleep-deprived individuals with the two-process model

We present a new method for developing individualized biomathematical models that predict performance impairment for individuals restricted to total sleep loss. The underlying formulation is based on the two-process model of sleep regulation, which has been extensively used to develop group-average models. However, in the proposed method, the parameters of the two-process model are systematically adjusted to account for an individual's uncertain initial state and unknown trait characteristics, resulting in individual-specific performance prediction models. The method establishes the initial estimates of the model parameters using a set of past performance observations, after which the parameters are adjusted as each new observation becomes available. Moreover, by transforming the nonlinear optimization problem of finding the best estimates of the two-process model parameters into a set of linear optimization problems, the proposed method yields unique parameter estimates. Two distinct data sets are used to evaluate the proposed method. Results of simulated data (with superimposed noise) show that the model parameters asymptotically converge to their true values and the model prediction accuracy improves as the number of performance observations increases and the amount of noise in the data decreases. Results of a laboratory study (82 h of total sleep loss), for three sleep-loss phenotypes, suggest that individualized models are consistently more accurate than group-average models, yielding as much as a threefold reduction in prediction errors. In addition, we show that the two-process model of sleep regulation is capable of representing performance data only when the proposed individualized model is used.

Optimal designs for studying bioimpedance

D-optimal designs for nonlinear fixed and mixed effects models are explored, and the theory is applied to the measurement and analysis of bioelectrical impedance. Bioimpedance is known to vary more at extreme frequencies than others. D-optimal designs that account for this variation, and also possible mis-specification of initial parameter estimates, are considered in an attempt to find designs that will provide good parameter estimates in practice.

Retrieving soil temperature profile by assimilating MODIS LST products with ensemble Kalman filter

Proper estimation of initial state variables and model parameters are vital importance for determining the accuracy of numerical model prediction. In this work, we develop a one-dimensional land data assimilation scheme based on ensemble Kalman filter and Common Land Model version 3.0 (CoLM). This scheme is used to improve the estimation of soil temperature profile. The leaf area index (LAI) is also updated dynamically by MODIS LAI production and the MODIS land surface temperature (LST) products are assimilated into CoLM. The scheme was tested and validated by observations from four automatic weather stations (BTS, DRS, MGS, and DGS) in Mongolian Reference Site of CEOP during the period of October 1, 2002 to September 30, 2003. Results indicate that data assimilation improves the estimation of soil temperature profile about 1 K. In comparison with simulation, the assimilation results of soil heat fluxes also have much improvement about 13 W m − 2 at BTS and DGS and 2 W m − 2 at DRS and MGS, respectively. In addition, assimilation of MODIS land products into land surface model is a practical and effective way to improve the estimation of land surface variables and fluxes.

Determination of pattern centre in EBSD using the moving‐screen technique

The 'moving-screen' or 'pattern magnification' method of calibration for electron backscatter diffraction (EBSD) was reformulated to develop a high-precision technique requiring no crystallographic knowledge of the specimen and no initial estimates of the calibration parameters. The technique depends upon the accurate displacement of the screen and camera assembly. Corresponding points are selected, interactively, from EBSD patterns. It is suggested that, as an alternative, the selection of points from the Hough transform could lead to a completely automated routine.

Constitutive modelling of viscoplasticity in a nickel-based superalloy at high temperature

Inelastic deformation behaviour of an advanced nickel-base superalloy has been investigated at 650 °C under multiple step cyclic and dwell loading conditions and for different strain rates. A viscoplastic constitutive model with strain range dependent cyclic hardening has been proposed. The model is based on the unified Chaboche model modified to account for non-linear kinematic hardening and plastic strain memory. For uniaxial loading, the initial estimates of material parameters were derived from simple strain controlled tests whilst a optimization procedure has been used to obtain the final material parameters. In this paper description of the proposed model, interpretation of the parameters and the optimisation procedure are presented. The predictions of the model are compared with the results for dwell loading tests with different strain rates, cyclic loading tests with selected strain ranges and creep tests under three stress levels. The results predictions show notable improvement compared to the predictions from previous work with respect to stress amplitude. The correlation coefficients and sensitivities of the optimised material parameters were also examined, in justification for the new model.

An Easy Calibration Method for Central Catadioptric Cameras

Abstract Central catadioptric cameras are widely used in virtual reality and robot navigation, and the camera calibration is a prerequisite for these applications. In this paper, we propose an easy calibration method for central catadioptric cameras with a 2D calibration pattern. Firstly, the bounding ellipse of the catadioptric image and field of view (FOV) are used to obtain the initial estimation of the intrinsic parameters. Then, the explicit relationship between the central catadioptric and the pinhole model is used to initialize the extrinsic parameters. Finally, the intrinsic and extrinsic parameters are refined by nonlinear optimization. The proposed method does not need any fitting of partial visible conic, and the projected images of 2D calibration pattern can easily cover the whole image, so our method is easy and robust. Experiments with simulated data as well as real images show the satisfactory performance of our proposed calibration method.

A New Method for Measurement and Reduction of Software Complexity

This paper develops an improved structural software complexity metrics named information flow complexity which is closely related to the reliability of software. Together with the three software complexity metrics, the total software complexity is measured and some rules to reduce the complexity are presented in the paper. To illustrate and explain the process of measurement and reduction of software complexity, several examples and experiments are given. It is proposed that software complexity metrics can be measured earlier in software development and can provide substantial information of software systems whose reliability can be modeled and used in the determination of initial parameter estimation.

Fraction of design space plots for generalized linear models

The use of graphical methods for comparing the quality of prediction throughout the design space of an experiment has been explored extensively for responses modeled with standard linear models. In this paper, fraction of design space (FDS) plots are adapted to evaluate designs for generalized linear models (GLMs). Since the quality of designs for GLMs depends on the model parameters, initial parameter estimates need to be provided by the experimenter. Consequently, an important question to consider is the design's robustness to user misspecification of the initial parameter estimates. FDS plots provide a graphical way of assessing the relative merits of different designs under a variety of types of parameter misspecification. Examples using logistic and Poisson regression models with their canonical links are used to demonstrate the benefits of the FDS plots.

Effect of Additives on Shape Evolution during Electrodeposition: I. Multiscale Simulation with Dynamically Coupled Kinetic Monte Carlo and Moving-Boundry Finite-Volume Codes

A multiscale simulation model was developed to simulate shape evolution during copper electrodeposition in the presence of additives. The model dynamically coupled a kinetic Monte Carlo (KMC) model (for surface chemistry and roughness evolution) with a finite volume (FV) model (for transport and chemical reactions in the electrolyte) and a level-set code (for tracking macroscopic movement of the metal/electrolyte interface). The KMC code was coarse-grained and used a multisite mesoparticle approach to account for the adsorption of large molecules. A multilevel grid approach was used to achieve numerical efficiency and accurate interface tracking. The model was demonstrated for an application involving in-fill of two-dimensional trenches by copper electrodeposition with additives [bis(3-sulfopropyl)disulfide], polyethylene glycol, chloride, and l-(2-hydroxyethyl)-2-imidazolidinethione. Demonstration calculations were carried out with use of initial estimates for the physicochemical parameters associated with a proposed reaction mechanism. The approach was found to be feasible for computing stable, dynamic behavior during macroscopic shape evolution over extended periods of time while simultaneously tracking microscopic roughness evolution associated with nearly molecular scale events at the surface. Numerical results include predictions of the surface concentration distributions as a function of time and distance for each reactant, product, and intermediate species associated with the proposed reaction mechanism.

Model of the evolution of nanoparticles to crystals via an aggregative growth mechanism

A mechanistic mathematical model is presented that describes the evolution of precursor nanoparticles (PN) to crystals. The model treats PN as pseudospecies evolving through an arbitrary number of reversible first-order steps and having the ability to contribute to growth by aggregation with existing crystals after they evolve beyond a certain stage. The number concentration of any of the intermediate PN is small compared to the initial number concentration of PN. If quasi-steady-state (QSS) is assumed for these intermediate PN, the model can be solved analytically. The analytical solution can be used to obtain initial parameter estimates for the full, dynamic model. DLVO interactions between coalescing PN are accounted for in the model through the coalescence kernel, which is shown to be very sensitive to the size of the particles and to the surface potential of the particles. Modifications are made to the mechanism to increase the likelihood of older PN to coalesce and to decrease the likelihood of older PN to dissolve. The maxima in the simulated crystal size distributions (CSD) propagate to larger sizes over time in a propagating front manner and, in some cases, a peaked population can form.