Sequential Estimation of the Parameters in a Trigonometric Regression Model with the Gaussian Coloured Noise

This paper considers the estimation problem for a trigonometric regression model with the noise specified by the Ornstein–Uhlenbeck process with unknown parameter. We propose a sequential procedure which ensures a prescribed mean square precision uniformly in the nuisance parameter. The asymptotic behaviour of the procedure duration mean has been studied. This revised version was published online in August 2006 with corrections to the Cover Date.     

A first-principles study of surface and subsurface H on and in Ni(1 1 1): diffusional properties and coverage-dependent behavior

Periodic, self-consistent, density functional theory (GGA-PW91) calculations are performed for both surface and subsurface atomic hydrogen on and in Ni(1 1 1). At a low coverage (θ=0.25 ML), the binding energies (BEs) of a hydrogen atom in surface fcc, subsurface octahedral (first layer), and subsurface octahedral (second layer) sites are −2.89, −2.18, and −2.11 eV, respectively. The activation energy barriers for hydrogen diffusion from the surface to the first subsurface layer and from the first to the second subsurface layer are estimated to be 0.88 and 0.52 eV, respectively. In the entire coverage range studied, hydrogen occupies surface fcc and hcp sites and subsurface octahedral sites. In addition, the magnitude of the BE per hydrogen atom and the magnetization of the nickel slabs both decrease as hydrogen coverage increases. Vibrational frequencies of hydrogen at various surface and subsurface sites are calculated and are in reasonable agreement with experimental data. A phase stability calculation with a 2 × 2 surface unit cell shows that a p(2 × 2)-2H overlayer structure (θ=0.5 ML) and a p(1 × 1)-1H structure (θ=1.0 ML) are stable at low hydrogen pressures, in agreement with numerous experimental results. A very large increase in pressure is required to populate subsurface sites. After such an increase occurs, the first subsurface layer is filled completely.     

HREELS signal processing via wavelets

In spectroscopy, the recorded spectra can often be modelled as the noisy convolution product of an instrumental function with the ‘true’ signal to be estimated. Such models have often been used for high‐resolution electron energy‐loss spectroscopy (HREELS). In this article, a new method is suggested to estimate the ‘true’ HREELS signal, i.e. the original electronic diffusion function with ‘true’ peak intensities. Our method relies upon the use of wavelets that, because they exhibit simultaneous time and frequency localization, are well‐suited for signal analysis. Firstly, a wavelet shrinkage algorithm is used to filter the noise. This is achieved by decomposing the noisy signal into an appropriate wavelet basis and then thresholding the wavelet coefficients that contain noise. This algorithm has a particular threshold related to frequency and time. Secondly, the broadening due to the instrumental response is eliminated through a deconvolution process. This step mainly rests on the existing relation between the Lipschitz regularity of the signal and the decay with scale of its wavelet coefficients and on least squares. The efficiency of this technique is highlighted by comparing the results obtained with those provided by other published methods. This work is the second in a series of three papers in this issue. The first one presents background knowledge on the wavelets required to understand the estimation methods. The third paper explores the application of wavelet filtering and deconvolution techniques to x‐ray photoelectron spectroscopy. Copyright © 2004 John Wiley & Sons, Ltd.     

Relative intensity noise of mode-locked fiber grating external cavity semiconductor lasers

The noise of hybrid soliton pulse source (HSPS) with linearly chirped Gaussian apodized fiber Bragg grating is analyzed by couple-mode equations including spontaneous emission noise when the HSPS is mode-locked. Relative intensity noise is calculated using numerical solutions of these equations. It is shown that transform limited pulses are generated over a wide tuning range around the fundamental mode-locking frequency with low spontaneous noise. However, a high noise level affects the operation of device, and therefore transform-limited pulses are not obtainable over a wide tuning range. It is also shown that noise is extremely sensitive to the RF and DC currents, linewidth enhancement factor, gain saturation parameter and spontaneous coupling factor.     

Behavior of zirconium in the ZrOW(100) system at high temperature, studied by ISS, AES and work-function measurements

The behavior of zirconium atoms at the W(100) surface associated with oxygen adsorption at different sample temperatures has been studied by Auger electron spectroscopy (AES), ion scattering spectroscopy (ISS), and the relative change of the work function (Δф) measured by the onset of the secondary electron energy distribution. The results have revealed: (i) adsorption of zirconium onto the W(100) surface followed by the elevation of the sample temperature up to 1710 K in an oxygen partial pressure of 2.7 × 10⁻⁴ induces complete diffusion of zirconium atoms into the W(100) substrate; (ii) further exposure of oxygen induces co-existence of oxygen and tungsten on the surface at 1710 K, resulting in a work function of 4.37 eV; (iii) keeping the sample temperature at 1710 K, simple evacuation of the system has resulted in surface segregation of zirconium atoms to the surface to form a zirconium atomic layer on the top-most surface, reducing the work function to 2.7 eV. The results have revealed that this specific behavior of zirconium atoms at high temperature assures, with very good reproducibility, the highly stable performance and long service life of Zr---O/W(100)-emitters in practical use, even in a low vacuum of 10⁻⁶ Pa.     

奇偶相干态的Wehrl熵和Shannon熵

研究了奇、偶相干态的Ｗｅｈｒｌ熵和Ｓｈａｎｎｏｎ熵，结果表明，其Ｗｅｈｒｌ熵随总噪声的增加而增大，并趋向同一个常数。Ｓｈａｎｎｏｎ熵随力学量起伏的减小而减小，且这两种态的Ｓｈａｎｎｏｎ熵均可小于相干态的Ｓｈａｎｎｏｎ熵，但这时并不与压缩效应的出现相对应，因而Ｓｈａｎｎｏｎ熵不能成为压缩存在的判据。     

调制不稳定性对级联放大光纤传输系统信噪比的影响

在考虑光纤损耗和级联放大器的情况下，推导了调制不稳定性的产生条件和增益的普适解析表达式，分析了调制不稳定性对信噪比的影响，给出了一个新的计算信噪比的表达式。     

Flicker noise in degenerately doped Si single crystals near the metal-insulator transition

In this paper we report some of the important results of experimental investigations of the flicker noise near the metal-insulator (MI) transition in doped silicon single crystals. This is the first comprehensive work to study low-frequency noise in heavily doped Si over an extensive temperature range (2 K<T<500 K). The measurements of conductance fluctuations (flicker noise) were carried out in the frequency range 10⁻²<f<4 × 10¹ Hz in single crystalline Si across the MI transition by doping with phosphorous and boron. The magnitude of noise in heavily doped Si is much larger than that seen in lightly doped Si over the whole temperature range. The extensive temperature range covered allowed us to detect two distinct noise mechanisms. At low temperatures (T<100 K) universal conductance fluctuations (UCF) dominate and the spectral dependence of the noise is determined by dephasing the electron from defects with two-levels (TLS). At higher temperatures (T>200 K) the noise arises from activated defect dynamics. As the MI transition is approached, the 1/f spectral power, typical of the metallic regime, gets modified by the presence of discrete Lorentzians which arise from generation-recombination process which is the characteristic of a semiconductor.     

Noise-induced asymptotic periodicity in a piecewise linear map

We examine asymptotically periodic density evolution in one-dimensional maps perturbed by noise, associating the macroscopic state of these dynamical systems with a phase space density. For asymptotically periodic systems density evolution becomes periodic in time, as do some macroscopic properties calculated from them. The general formalism of asymptotic periodicity is examined and used to calculate time correlations along trajectories of these maps as well as their limiting conditional entropy. The time correlation is shown to naturally decouple into periodic and stochastic components. Finally, asymptotic periodicity is studied in a noise-perturbed piecewise linear map, focusing on how the variation of noise amplitude can cause a transition from asymptotic periodicity to asymptotic stability in the density evolution of this system.     

Vowel Intelligibility in Classical Singing

Vowel intelligibility during singing is an important aspect of communication during performance. The intelligibility of isolated vowels sung by Western classically trained singers has been found to be relatively low, in fact, decreasing as pitch rises, and it is lower for women than for men. The lack of contextual cues significantly deteriorates vowel intelligibility. It was postulated in this study that the reduced intelligibility of isolated sung vowels may be partly from the vowels used by the singers in their daily vocalises. More specifically, if classically trained singers sang only a few American English vowels during their vocalises, their intelligibility for American English vowels would be less than for those classically trained singers who usually vocalize on most American English vowels. In this study, there were 21 subjects (15 women, 6 men), all Western classically trained performers as well as teachers of classical singing. They sang 11 words containing 11 different American English vowels, singing on two pitches a musical fifth apart. Subjects were divided into two groups, those who normally vocalize on 4, 5, or 6 vowels, and those who sing all 11 vowels during their daily vocalises. The sung words were cropped to isolate the vowels, and listening tapes were created. Two listening groups, four singing teachers and five speech-language pathologists, were asked to identify the vowels intended by the singers. Results suggest that singing fewer vowels during daily vocalises does not decrease intelligibility compared with singing the 11 American English vowels. Also, in general, vowel intelligibility was lower with the higher pitch, and vowels sung by the women were less intelligible than those sung by the men. Identification accuracy was about the same for the singing teacher listeners and the speech-language pathologist listeners except for the lower pitch, where the singing teachers were more accurate.