Home About us Contact
|
Home About us Contact | ||
|
|
||
Least-squares Estimation (least-square + estimation)
Selected AbstractsA pattern-based approach for multiple removal applied to a 3D Gulf of Mexico data setGEOPHYSICAL PROSPECTING, Issue 2 2006Antoine Guitton ABSTRACT Surface-related multiples are attenuated for one sail line and one streamer of a 3D data set (courtesy of Compagnie Générale de Géophysique). The survey was carried out in the Gulf of Mexico in the Green Canyon area where salt intrusions close to the water-bottom are present. Because of the complexity of the subsurface, a wavefield method incorporating the full 3D volume of the data for multiple removal is necessary. This method comprises modelling of the multiples, where the data are used as a prediction operator, and a subtraction step, where the model of the multiples is adaptively removed from the data with matching filters. The accuracy of the multiple model depends on the source/receiver coverage at the surface. When this coverage is not dense enough, the multiple model contains errors that make successful subtraction more difficult. In these circumstances, one can either (1) improve the modelling step by interpolating the missing traces, (2) improve the subtraction step by designing methods that are less sensitive to modelling errors, or (3) both. For this data set, the second option is investigated by predicting the multiples in a 2D sense (as opposed to 3D) and performing the subtraction with a pattern-based approach. Because some traces and shots are missing for the 2D prediction, the data are interpolated in the in-line direction using a hyperbolic Radon transform with and without sparseness constraints. The interpolation with a sparseness constraint yields the best multiple model. For the subtraction, the pattern-based technique is compared with a more standard, adaptive-subtraction scheme. The pattern-based approach is based on the estimation of 3D prediction-error filters for the primaries and the multiples, followed by a least-squares estimation of the primaries. Both methods are compared before and after prestack depth migration. These results suggest that, when the multiple model is not accurate, the pattern-based method is more effective than adaptive subtraction at removing surface-related multiples while preserving the primaries. [source] Optimality for the linear quadratic non-Gaussian problem via the asymmetric Kalman filterINTERNATIONAL JOURNAL OF ADAPTIVE CONTROL AND SIGNAL PROCESSING, Issue 1 2004Rosario Romera Abstract In the linear non-Gaussian case, the classical solution of the linear quadratic Gaussian (LQG) control problem is known to provide the best solution in the class of linear transformations of the plant output if optimality refers to classical least-squares minimization criteria. In this paper, the adaptive linear quadratic control problem is solved with optimality based on asymmetric least-squares approach, which includes least-squares criteria as a special case. Our main result gives explicit solutions for this optimal quadratic control problem for partially observable dynamic linear systems with asymmetric observation errors. The main difficulty is to find the optimal state estimate. For this purpose, an asymmetric version of the Kalman filter based on asymmetric least-squares estimation is used. We illustrate the applicability of our approach with numerical results. Copyright © 2004 John Wiley & Sons, Ltd. [source] All linear methods are equal,and extendible to (some) nonlinearitiesINTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL, Issue 8 2002Janos GertlerArticle first published online: 17 JUN 200 Abstract Several linear methods of residual generation for fault detection and diagnosis are reviewed. The parity relation approach is introduced in some detail, for both additive and parametric faults. The Chow,Willsky scheme, various diagnostic observers and principal component analysis are compared to the additive version. The ,local approach' and the least-squares estimation of parameter changes are shown to be related to the parametric variant. Nonlinear extensions are demonstrated for all the techniques under additive faults. Copyright © 2002 John Wiley & Sons, Ltd. [source] Generalized k -space decomposition with chemical shift correction for non-cartesian water-fat imagingMAGNETIC RESONANCE IN MEDICINE, Issue 5 2008Ethan K. Brodsky Abstract Chemical-shift artifacts associated with non-Cartesian imaging are more complex to model and less clinically acceptable than the bulk fat shift that occurs with conventional spin-warp Cartesian imaging. A novel k -space based iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL) approach is introduced that decomposes multiple species while simultaneously correcting distortion of off-resonant species. The new signal model accounts for the additional phase accumulated by off-resonant spins at each point in the k -space acquisition trajectory. This phase can then be corrected by adjusting the decomposition matrix for each k -space point during the final IDEAL processing step with little increase in reconstruction time. The technique is demonstrated with water-fat decomposition using projection reconstruction (PR)/radial, spiral, and Cartesian spin-warp imaging of phantoms and human subjects, in each case achieving substantial correction of chemical-shift artifacts. Simulations of the point-spread-function (PSF) for off-resonant spins are examined to show the nature of the chemical-shift distortion for each acquisition. Also introduced is an approach to improve the signal model for species which have multiple resonant peaks. Many chemical species, including fat, have multiple resonant peaks, although such species are often approximated as a single peak. The improved multipeak decomposition is demonstrated with water-fat imaging, showing a substantial improvement in water-fat separation. Magn Reson Med 59:1151,1164, 2008. © 2008 Wiley-Liss, Inc. [source] Light-curve inversions with truncated least-squares principal components: Tests and application to HD 291095 = V1355 OrionisASTRONOMISCHE NACHRICHTEN, Issue 4 2008I.S. Savanov Abstract We present a new inversion code that reconstructs the stellar surface spot configuration from the light curve of a rotating star. Our code employs a method that uses the truncated least-squares estimation of the inverse problem's objects principal components. We use spot filling factors as the unknown objects. Various test cases that represent a rapidly-rotating K subgiant are used for the forward problem. Tests are then performed to recover the artificial input map and include data errors and input-parameter errors. We demonstrate the robustness of the solution to false input parameters like photospheric temperature, spot temperature, gravity, inclination, unspotted brightness and different spot distributions and we also demonstrate the insensitivity of the solution to spot latitude. Tests with spots peppered over the entire stellar surface or with phase gaps do not produce fake active longitudes. The code is then applied to ten years of V and I -band light curve data of the spotted sub-giant HD291095. A total of 22 light curves is presented. We find that for most of the time its spots were grouped around two active longitudes separated on average by 180°. Switches of the dominant active region between these two longitudes likely occurred about every 3.15±0.23 years while the amplitude modulation of the brightness occurred with a possible period of 3.0±0.15 years. For the first time, we found evidence that the times of the activity flips coincide with times of minimum light as well as minimum photometric amplitude, i.e. maximum spottedness. From a comparison with simultaneous Doppler images we conclude that the activity flips likely take place near the rotational pole of the star. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim) [source] | ||||||||||