Search

We present an analysis of the bandwidth of an asymmetric 8-beam Fourier grating as the beam multiplexer for a 4.7 THz local oscillator used in a heterodyne receiver. We take the grating designed for NASA GUSTO balloon observatory as an example to address the bandwidth question although it does not need to operate over a wide frequency range. By illuminating the grating at different frequencies from 4.445 to 5.045 THz, we simulated the changes of its performance in three aspects using COMSOL Multiphysics: diffraction efficiency, power uniformity, and the angular distribution of the output beams. These parameters can affect the coupling efficiency between the output beams of the grating and the beams of a mixer array. The bandwidth of the grating is found to be 230 GHz, corresponding to 4.9% of the operating frequency, which is sufficient for many applications.

Climate change has considerably increased the likelihood of experiencing extreme weather events. Governments in developing countries have a limited capacity to smooth the losses created by extreme weather, and could potentially benefit from the introduction of disaster funds, that is, ex-ante budgeting allocations for post-disaster reconstruction. So far the implementation of disaster funds has been limited, in part because it is still unclear whether disaster funds provide a cost-effective way of coping with these losses. By taking advantage of the sharp rules that govern the municipal-level eligibility for reconstruction funds in Mexico, this paper provides some of the first estimates of the impact of disaster funds on local economic activity. The main finding is that access to disaster funding boosts local economic activity between 2 and 4 percent in the year following the disaster. Another finding is that the positive impact of disaster funds on local economic recovery can persist for as long as a year and a half after the disaster. Consistent with these findings, we additionally show that access to disaster funding leads to a large and sustained 76 percent increase in the growth of local construction employment. This labor market impact slightly precedes the overall increase in local economic activity.

This paper documents new facts about the joint evolution of firm performance and prices in international markets and proposes a theory of firm dynamics emphasizing the interaction between learning about demand and quality choice to explain the observed patterns. Using data from the Portuguese manufacturing sector, the paper documents that: (1) within narrow product categories, firms with longer spells of activity in export destinations tend to ship larger quantities at similar prices, thus obtaining larger export revenue; (2) older exporters tend to import more expensive inputs; and (3) revenue growth within destinations (conditional on initial size) tends to decline with market experience. The authors develop a model of endogenous input and output quality choices in a learning environment that is able to quantitatively account for these patterns. Counterfactual simulations reveal that minimum quality standards on exports reduce welfare by lowering entry in export markets and reallocating resources from old and large toward young and small firms.

A recent survey of rigorous impact evaluations of programs to help small and medium-size firms to formalize indicates that the programs do not seem to work for most informal firms. One of the few exceptions finds large effects of a tax simplification program in Brazil called SIMPLES on firms' formalization rates and performance indicators. Using the same data set but a different identification strategy, another study concludes that the program had limited effect on formalization rates. The aim of this paper is twofold. First, it revisits the two studies to reconcile their conflicting conclusions. Second, it investigates the validity of the identification strategy of both studies. The findings suggest that the conflicting results between the two studies are caused by the dates each used to identify when the program was put into effect. A robustness check indicates that data heaping and seasonality around November cast doubts on the identification strategy used in both studies to estimate the effect of this particular program.

The nuclear magnetic resonance of paramagnetic solids is usually characterized by the presence of large chemical shifts and shift anisotropies due to hyperfine interactions. Frequently the resulting spectra cover a frequency range of several megahertz, which is greater than the bandwidth of commercially available radio-frequency (RF) probes, making it impossible to acquire the whole spectrum in a single experiment. In these cases it common to record a series of spectra, in which the probe is tuned to a different frequency for each, and then sum the results to give the ``true'' spectrum. While this method is very widely used on static samples, the application of frequency stepping under magic-angle spinning (MAS) is less common, owing to the increased complexity of the spin dynamics when describing the interplay of the RF irradiation with the mechanical rotation of the shift tensor. In this paper, we present a theoretical description, based on the jolting frame formalism of Caravatti et al. [J. Magn. Reson. 55, 88 (1983)], for describing the spin dynamics of a powder sample under MAS when subjected to a selective pulse of low RF-field amplitude. The formalism is used to describe the frequency stepping method under MAS, and under what circumstances the true spectrum is reproduced. We also present an experimental validation of the methodology under ultra-fast MAS with the paramagnetic materials LiMnPO4 and TbCsDPA. (C) 2013 American Institute of Physics. [http://dx.doi.org/10.1063/1.4795001]

Proceedings of Interspeech 2011, Florence (Italy), We describe a new approach to automatic speaker recognition based in explicit modeling of temporal contours in linguistic units (TCLU). Inspired in successful work in forensic speaker identification, we extend the approach to design a fully automatic system, with a high potential for combination with spectral systems. Using SRI's Decipher phone, word and syllabic labels, we have tested up to 468 unit-based subsystems from 6 groups of lexically-determined units, namely phones, diphones, triphones, center phone in triphones, syllables and words, subsystems being combined at the score level. Evaluating with NIST SRE04 English-only 1s1s, their hierarchical fusion gives an EER of 4.20% (minDCF=0.018) from automatic formant tracking of conversational telephone speech. Combining extremely well with a Joint Factor Analysis system (from JFA EER of 4.25% to 2.47%, minDCF from 0.020 to 0.012), extensions as more robust prosodic or spectral features are likely to further improve this approach., This work has been supported by MEC research stay grant PR-2010-123, MICINN project TEC09-14179, ForBayes project CCG10-UAM/TIC-5792 and Catedra UAM-Telefonica.

Generalized additive models (GAM) are multivariate nonparametric regressions for non-Gaussian responses including binary and count data. We propose a spline-backfitted kernel (SBK) estimator for the component functions. Our results are for weakly dependent data and we prove oracle efficiency. The SBK techniques is both computational expedient and theoretically reliable, thus usable for analyzing high-dimensional time series. Inference can be made on component functions based on asymptotic normality. Simulation evidence strongly corroborates with the asymptotic theory.

Generalized additive models (GAM) are multivariate nonparametric regressions for non-Gaussian responses including binary and count data. We propose a spline-backfitted kernel (SBK) estimator for the component functions. Our results are for weakly dependent data and we prove oracle efficiency. The SBK techniques is both computational expedient and theoretically reliable, thus usable for analyzing high-dimensional time series. Inference can be made on component functions based on asymptotic normality. Simulation evidence strongly corroborates with the asymptotic theory.

Application of nonparametric and semiparametric regression techniques to high-dimensional time series data has been hampered due to the lack of effective tools to address the ``curse of dimensionality.'' Under rather weak conditions, we propose spline-backfitted kernel estimators of the component functions for the nonlinear additive time series data that are both computationally expedient so they are usable for analyzing very high-dimensional time series, and theoretically reliable so inference can be made on the component functions with confidence. Simulation experiments have provided strong evidence that corroborates the asymptotic theory., Comment: Published in at http://dx.doi.org/10.1214/009053607000000488 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

Processing of broadband signals induced on an antenna array using a tapped delay line filter and a set of steering delays has two problems. Firstly one needs to manipulate large matrices to estimate the filter coefficients. Secondly the use of steering delays is not only cumbersome but implementation errors cause loss of system performance. This thesis looks at both of these problems and presents elegant solutions by developing and studying a design method referred to as the DFT method, which does not require steering delays and is computationally less demanding compared to existing methods. Specifically the thesis studies and compares the performance of a time domain element space beamformer using the proposed method and that using an existing method, and develops the DFT method when the processor is implemented in partitioned form. The study presented in the thesis shows that the processors using the DFT method are robust to look direction errors and require less computation than that using the existing method for comparable performance. The thesis further introduces a broadband beamformer design which does not require any steering delays between the sensors and the tapped delay line section as is presently the case. It has the capability of steering the array in an arbitrary direction with a specified frequency response in the look direction while canceling unwanted uncorrelated interferences. The thesis presents and compares the performance of a number of techniques to synthesize an antenna pattern of a broadband array. These techniques are designed to produce isolated point nulls as well as broad sector nulls and to eliminate the need for the steering delays. Two of the pattern synthesis techniques presented in the thesis allow optimization against unwanted interferences in unknown directions. The techniques allow formulation of a beamforming problem such that the processor is not only able to place nulls in specified directions but also able to cancel directional interferences in unknown directions along with a specified frequency response in the look direction over a band of interest. The thesis also presents a set of directional constraints such that one does not need steering delays and an array can be constrained in an arbitrary direction with a specified frequency response. The constraints presented in the thesis are simple to implement. Based on these constraints a pattern synthesis technique for broadband antenna array is also presented.

Perceiving motion patterns in visual scenes in which speed or motion direction varies over space while average luminance remains constant presents a processing task that requires at least two separate stages of neural spatio-temporal filtering. We have previously probed the transfer of information between these stages of filtering identifying a largely scale invariant process in which narrowband initial motion sensitive filters are coupled with a broad range of spatial frequencies of secondary filters, with an optimal coupling – in terms of optimal observer visual sensitivity – at a frequency ratio of around twelve. In the current work, we used the same stimulus to investigate the possible presence of multiple secondary filtering mechanisms and their associated bandwidths. Using a forced choice psychophysical task with both a detection and an identification component, we presented experimental blocks containing stimuli with one of two different modulator frequencies in each trial to measure the frequency difference at which the detection performance matched the identification of the frequency. We found that at a frequency differences of about 2.2 octaves, performance of both tasks was similar, and the processing could therefore be inferred to occur in independent frequency channels. The same observation was confirmed for stimuli presented at a longer viewing distance. We conclude that for the motion gradient stimuli, there are secondary filtering mechanisms with a moderately broad bandwidth of over 2 octaves that underlie our sensitivity for detecting motion gradients of different modulation frequency. These are likely to be implemented at least in part within the dorsal stream of extra-striate cortex.