Browsing by Author "Guarini, M."

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Hybrid FEM/BEM modeling of finite-sized photonic crystals for semiconductor laser beams
    (WILEY, 2010) Jerez Hanckes, C.; Duran, M.; Guarini, M.
    We propose a 2-D finite element/boundary element hybrid method for calculating the spatial distribution and frequency response of electromagnetic waves coming from a semiconductor laser when interacting with a finite-sized photonic crystal. We thus provide a flexible tool for the design of novel optical and microwave devices, among other applications. In opposition to current methodologies, we simultaneously take into account the laser modes, the finiteness of the crystal, and the unboundedness of the isotropic medium in which the crystal is embedded. At the laser output, instead of approximating reflected and transmitted beams by plane waves, we use the more realistic Hermite-Gauss functions. In the isotropic medium, we set an artificial boundary encircling the crystal and define exterior and interior domains. Radiating solutions for the scattered far field over the exterior are derived analytically through a series of Hankel polynomials. The interior domain is described by a finite element formulation coupled with Dirichlet-to-Neumann maps enforcing laser and far-field behaviors. Results and error analyses are provided in view of future improvements. Copyright (C) 2010 John Wiley & Sons, Ltd.
  • Thumbnail Image
    Item
    Undersampling k-space using fast progressive 3D trajectories
    (WILEY, 2005) Spiniak, J.; Guesalaga, A.; Mir, R.; Guarini, M.; Irarrazaval, P.
    In 3D MRI, sampling k-space with traditional trajectories can be excessively time-consuming. Fast imaging trajectories are used in an attempt to efficiently cover the k-space and reduce the scan time without significantly affecting the image quality. In many applications, further reductions in scan time can be achieved via undersampling of the k-space; however, no clearly optimal method exists. In most 3D trajectories the k-space is divided into regions that are sampled with shots that share a common geometry (e.g., spirals). A different approach is to design trajectories that gradually but uniformly cover the k-space. In the current work, successive shots progressively add sampled regions to the 3D frequency space. By cutting the sequence short, a natural undersampled method is obtained. This can be particularly efficient because in these types of trajectories the contribution of new information by later shots is less significant. In this work the performance of progressive trajectories for different degrees of undersampling is assessed with trajectories based on missile guidance (MG) ideas. The results show that the approach can be efficient in terms of reducing the scan time, and performs better than the stack of spirals (SOS) technique, particularly under nonideal conditions.