<pre>The Average Transform Smooth (ATS) statistical methods [McRae, Mallows, and Cleveland], are applied to measurements of a non-gaussian random variable to make them close to gaussian. This gaussianization makes use of the well known concept of variance stabilizing transformation, but takes it further by first averaging blocks of r measurements, transforming next, and then smoothing. The smoothing can be nonparametric, or can be the fitting of a parametric model. The gaussianization makes analysis simpler and more effective.</pre><pre><br></pre><pre>In this work ATS is applied to the periodogram of a stationary parametric time series, and makes use of the periodogram large sample properties given the true power spectrum [Brillinger], to develop a new approach to parametric time series model estimation and model diagnostics. The ATS results and the theory are reformulated as a regression model, PPS-REG, involving true power spectrum and the periodogram. PPS-REG has attractive properties: iid gaussian error terms with mean 0 and a known variance; accurate estimation; much faster estimation than the classical maximum likelihood when the time series is large; enables the use of the very powerful classical regression model diagnostics; bases the diagnostics on the power spectrum, adding substantially to the standard use of the autocovariance function for diagnosing the fits of models specified in the time domain.</pre>

No description available.

No description available.

No description available.

The aim of this master’s thesis is to learn the problem of analysis and identification of significant spectral components in speech signal. Based on learning a special literature chooses the suitable methods of spectrum estimate. Does learning the literature in specification of testing of spectral components significate. Makes a procedure for identification of chosen speech formants. Does this procedure for audio signals both of in stress and in normal state. Estimates the results, compares efficiency of chosen methods and determine threshold for chosen formant of analyzed stress signal. States the recommendations for speech spectral analysis in stress situation.

No / Abstract: The scarcity of frequency spectrum used for wireless communication systems has attracted a considerable amount of attention in recent years. The cognitive radio (CR) terminology has been widely accepted as a smart platform mainly aimed at the efficient interrogation and utilization of permitted spectrum. Non-parametric spectrum sensing, or estimation, represents one of the prominent tools that can be proposed when CR works under an undetermined environment. As such, the periodogram, filter bank, and multi-taper methods are well considered in many studies without relying on the transmission channel's characteristics. A unified approach to all these non-parametric spectrum sensing techniques is presented in this paper with analytical and performance comparison using simulation methods. Results show that the multi-taper method outperforms the others.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第22828号 / 医博第4667号 / 新制||医||1047(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 村井 俊哉, 教授 古川 壽亮, 教授 高橋 淳 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Point processes are common in many physical applications found in engineering and biology. These processes can be observed in one-dimension as a time series or two-dimensions as a spatial point pattern with extensive amounts of literature devoted to their analyses. However, if the observed process is a hybrid of spatial and temporal point process, very few practical methods exist. In such cases, practitioners often remove the temporal component and analyze the spatial dependencies. This marginal spatial analysis may lead to misleading results if time is an important factor in the process. In this dissertation we extend the current analysis of spatial point patterns to include a temporal dimension. First- and second-order intensity measures for analyzing spatiotemporal point patterns are explicitly defined. Estimation of first-order intensities are examined using 3-dimensional smoothing techniques. Conditions for weak stationarity are provided so that subsequent second-order analysis can be conducted. We consider second-order analysis of spatiotemporal point patterns first in the space-time domain through an extension of Ripley's Κ-function. An alternative analysis is given in the frequency domain though construction of a spatiotemporal periodogram. The methodology provided is tested through simulation of spatiotemporal point patterns and by analysis of a real data set. The biological application concerns the estimation of the homerange of groups of the endangered red-cockaded woodpecker in the Fort Bragg area of North Carolina. Monthly or bimonthly point patterns of the bird distribution are analyzed and integrated over a 23 month period. / Ph. D.

The use of ultra-wideband (UWB) signals holds great promise for remote monitoring of vital-signs which has applications in the medical, for first responder and in security. Previous research has shown the feasibility of a UWB-based radar system for respiratory and heartbeat rate estimation. Some simulation and real experimental results are presented to demonstrate the capability of the respiration rate detection. However, past analysis are mostly based upon the assumption of an ideal experiment environment. The accuracy of the estimation and interference factors of this technology has not been investigated. This thesis establishes an analytical framework for the FFT-based signal processing algorithms to detect periodic bio-signals from a single target. Based on both simulation and experimental data, three basic challenges are identified: (1) Small body movement during the measurement interval results in slow variations in the consecutive received waveforms which mask the signals of interest. (2) The relatively strong respiratory signal with its harmonics greatly impact the detection of heartbeat rate. (3) The non-stationary nature of bio-signals creates challenges for spectral analysis. Having identified these problems, adaptive signal processing techniques have been developed which effectively mitigate these problems. Specifically, an ellipse-fitting algorithm is adopted to track and compensate the aperiodic large-scale body motion, and a wavelet-based filter is applied for attenuating the interference caused by respiratory harmonics to accurately estimate the heartbeat frequency. Additionally, the spectrum estimation of non-stationary signals is examined using a different transform method. Results from simulation and experiments show that substantial improvement is obtained by the use of these techniques. Further, this thesis examines the possibility of multi-target detection based on the same measurement setup. Array processing techniques with subspace-based algorithms are applied to estimate multiple respiration rates from different targets. The combination of array processing and single- target detection techniques are developed to extract the heartbeat rates. The performance is examined via simulation and experimental results and the limitation of the current measurement setup is discussed. / Master of Science

Dans ce travail nous nous intéressons à l'estimation de la densité spectrale par la méthode du noyau pour des processus à temps continu et des champs aléatoires observés selon des schémas d'échantillonnage (ou plan d'expériences) discrets aléatoires. Deux types d'échantillonnage aléatoire sont ici considérés : schémas aléatoires dilatés, et schémas aléatoires poissonniens. Aucune condition de gaussiannité n'est imposée aux processus et champs étudiés, les hypothèses concerneront leurs cumulants.En premier nous examinons un échantillonnage aléatoire dilaté utilisé par Hall et Patil (1994) et plus récemment par Matsuda et Yajima (2009) pour l'estimation de la densité spectrale d'un champ gaussien. Nous établissons la convergence en moyenne quadratique dans un cadre plus large, ainsi que la vitesse de convergence de l'estimateur.Ensuite nous appliquons l'échantillonnage aléatoire poissonnien dans deux situations différentes : estimation spectrale d'un processus soumis à un changement de temps aléatoire (variation d'horloge ou gigue), et estimation spectrale d'un champ aléatoire sur R2. Le problème de l'estimation de la densité spectrale d'un processus soumis à un changement de temps est résolu par projection sur la base des vecteurs propres d'opérateurs intégraux définis à partir de la fonction caractéristique de l'accroissement du changement de temps aléatoire. Nous établissons la convergence en moyenne quadratique et le normalité asymptotique de deux estimateurs construits l'un à partir d'une observation continue, et l'autre à partir d'un échantillonnage poissonnien du processus résultant du changement de temps.La dernière partie de ce travail est consacrée au cas d'un champ aléatoire sur R2 observé selon un schéma basé sur deux processus de Poissons indépendants, un pour chaque axe de R2. Les résultats de convergence sont illustrés par des simulations / In this work, we are dealing in the kernel estimation of the spectral density for a continuous time process or random eld observed along random discrete sampling schemes. Here we consider two kind of sampling schemes : random dilated sampling schemes, and Poissonian sampling schemes. There is no gaussian condition for the process or the random eld, the hypotheses apply to their cumulants.First, we consider a dilated sampling scheme introduced by Hall and Patil (1994) and used more recently by Matsuda and Yajima (2009) for the estimation of the spectral density of a Gaussian random eld.We establish the quadratic mean convergence in our more general context, as well as the rate of convergence of the estimator.Next we apply the Poissonian sampling scheme to two different frameworks : to the spectral estimation for a process disturbed by a random clock change (or time jitter), and to the spectral estimation of a random field on R2.The problem of the estimatin of the spectral density of a process disturbed by a clock change is solved with projection on the basis of eigen-vectors of kernel integral operators defined from the characteristic function of the increment of the random clock change. We establish the convergence and the asymptotic normality of two estimators contructed, from a continuous time observation, and the other from a Poissonian sampling scheme observation of the clock changed process.The last part of this work is devoted to random fields on R2 observed along a sampling scheme based on two Poisson processes (one for each axis of R2). The convergence results are illustrated by some simulations