Wave propagation is a common phenomenon in aircraft structures resulting from high velocity transient loadings like bird hit, gust etc. Apart from understanding the behavior of structures under such loading, wave propagation analysis is also important to gain knowledge about their high frequency characteristics, which have several applications. The applications include structural health monitoring using diagnostic waves and control of wave transmission for reduction of noise and vibration. Transient loadings with high frequency content are associated with wave propagation. As a result, the higher modes of the structure participate in the response. Finite element (FE) modeling for such problem requires very fine mesh to capture these higher modes. This leads to large system size and hence large computational cost. Wave propagation problems are usually solved in frequency domain using fast Fourier transform (FFT) and spectral finite element method is one such technique which follows FE procedure in the transformed frequency domain. In this thesis, a novel wavelet based spectral finite element (WSFE) is developed for wave propagation analysis in finite dimension structures. In WSFE for 1-D waveguides, the partial differential wave equations are reduced to a set of ODEs using orthogonal compactly supported Daubechies scaling functions for temporal approximation. The localized nature of the Daubechies basis functions allows finite domain analysis and imposition of the boundary conditions. The reduced ODEs are usually solved exactly, the solution of which gives the dynamic shape functions. The interpolating functions used here are exact solution of the governing differential equation and hence, the exact elemental dynamic stiffness matrix is derived. Thus, In the absence of any discontinuities, one element is sufficient to model 1-D waveguide of any length. This elemental stiffness matrix can be assembled to obtain the global matrix as in FE and after solution, the time domain responses are obtained using the inverse wavelet transform. The developed technique circumvents several serious limitations of the conventional FFT based Spectral Finite Element (FSFE). In FSFE, the wave equations are reduced to ODEs using FFT for time approximation. The remaining part of the formulation is quite similar to that of WSFE. The required assumption of periodicity in FSFE, however, does not allow modeling of finite length structures. It results in “wrap around” problem, which distorts the response simulated using FSFE and a semi-infinite (“throw-off”) element is required for imparting artificial damping. This artificial damping occurs as the “throw off” element allows leakage of energy. In some cases, a very high damping can also be considered instead of “throw off” element to remove wrap around effects. In either cases, the damping introduced is much larger than any inherent damping that may be present in the structure. It should also be mentioned that even in presence of the artificial damping, a larger time window is required for removing the distortions completely. The developed WSFE method is completely free from such problems and can efficiently handle undamped finite length structures irrespective of the time window considered. Apart from this, FSFE allows imposition of only zero initial condition and in contrary any initial conditions can be used in WSFE. Though FSFE has problem in modeling finite length undamped structures for time domain analysis, it is well suited for performing frequency domain study of wave characteristics, namely, the determination of spectrum and dispersion relations. WSFE is also capable of extracting these frequency dependent wave properties, however only up to a certain fraction of the Nyquist frequency. This constraint results from the loss in frequency resolution due to the increase in time resolution in wavelet analysis, where the basis functions are bounded both in time and frequency. A price has to be paid in frequency domain in order to obtain a bound in the time domain. The consequence of this analysis is to impose a constraint on the time sampling rate for the simulation with WSFE, to avoid spurious dispersion. WSFE for 2-D waveguides are formulated using Daubechies scaling functions for both temporal and spatial approximations. The initial and boundary conditions, however, are imposed using two different methods, which are wavelet extrapolation technique and periodic extension or restraint matrix respectively. The 2-D WSFE is bounded in both the spatial directions unlike 2-D FSFE, which is essentially unbounded in one spatial direction. Apart from this, 2-D WSFE is also free from “wrap around” problem similar to 1-D WSFE due to the localized nature of the basis functions used for temporal approximation. In this thesis, WSFE is developed for isotropic 1-D and 2-D waveguides for time and frequency domain analysis. These include elementary rod, Euler-Bernoulli and Timoshenko beams in 1-D modeling, and plates and axisymmetric cylinders in 2-D modeling. The wave propagation responses simulated using WSFE for these waveguides are validated using FE results. The advantages of the proposed technique over the corresponding FSFE method are also highlighted all through the numerical examples. Next part of the thesis involves the extension of the developed WSFE technique for modeling composite and nano-composite structures to study their wave propagation behavior. Due to their anisotropic nature, analysis of composite structures, particularly high frequency transient analysis is much more complicated compared to the corresponding metallic structures. This is due to the presence of stiffness coupling in these structures. Superior mechanical properties of composites, however, are making them integral parts of an aircraft and thus they often experience such short duration, high velocity impact Loadings. Very few literatures report the response of composite structures subjected to such high frequency excitations. Here, WSFE is formulated for a higher order composite beam with axial, flexural, shear and contractional degrees of freedom. WSFE is also formulated for composite plates using classical laminated plate theory with axial and flexural degrees of freedom. Simulations performed using these WSFE models are used to study the higher order and elastic coupling effects on the wave propagation responses. Carbon nanotubes (CNTs) and their composites are attracting a great deal of experimental and theoretical research world-wide. The recent trend in the literature shows a great interest in the dynamic and wave characteristics of CNTs and nano-composites because of their several applications. In most of these applications, CNTs are used in the embedded form as it does not requires precise alignment of the nano-tubes. In addition, the extraordinary mechanical properties of CNTs are being exploited to achieve high strength nano-composite. Apart from the experimental studies and atomistic simulation to study the mechanical properties of CNTs and nano-composites, continuum modeling is also receiving much attention, mainly due to its computational viability. In this thesis, a 1-D WSFE is formulated for multi-wall carbon nanotube (MWNT) embedded composite modeled as beam using higher order layer-wise theory. This theory allows to model partial interfacial shear stress transfer, which normally occurs due to improper dispersion of CNTs in nano-composites. The effects of different matrix materials and fraction of shear stress transfer on the wave characteristics are studied. The responses obtained using other beam theories are also compared. The beam modeling does not allow capturing the radial motions of the CNT, which are important for several applications. These can be effectively captured by modeling the CNT using a 2-D axisymmetric model. Hence, a 2-D WSFE model is constructed to capture the high frequency characteristics of single-walled carbon nanotubes (SWNTs). The response of SWNT simulated using the developed model is validated with experimental and atomistic simulation results reported in the literature. The comparison are done for dispersion relation and also radial breathing mode frequencies. The effects of geometrical parameters, namely the radius and the wall thickness of the SWNT on the higher radial, longitudinal and coupled radial-longitudinal vibrational modes are analyzed. These behaviors are studied in both time and frequency domains. Such time domain analyses of finite length SWNT are not possible with the Fourier transform based techniques reported in literature, although, such analyses are important particularly for sensor applications of SWNT. Spectral finite element method is very much suited for solution of inverse problems like force reconstruction from the measured wave response. This is because the technique is based on the concept of transfer function between the displacements (output) and applied forces (input). In the present work, WSFE is implemented for identification of impact force from the wave propagation responses simulated with FE and used as surrogate experimental results. The results show that WSFE can accurately reconstruct the impulse load applied to 1-D waveguides which include rod, Euler-Bernoulli beam and connected 2-D frame, even with highly truncated response. This is unlike FSFE, where the accuracy of the identified force depends largely on the time window of the measured responses. The detection of damage from the wave propagation analysis is another class of inverse problems considered in this thesis and is of utmost importance in the area of aircraft structural health monitoring. Here, the detection scheme is based on arrival time of the waves reflected from the damage. A novel detection technique based on wavelet filtering is proposed here and it is shown to work efficiently even in the presence of noise in the measured wave responses. Detection of damage requires an efficient damage model to simulate the mode of structural failure. In this regard, two spectrally formulated wavelet elements are proposed, one to model isotropic beam with through-width notch and the second to model composite beam with embedded de-lamination. In the first case, the response of the damaged beam is considered as the perturbation of the undamaged response and the linear perturbation analysis leads to a completely new set of dynamic stiffness matrix. In the second case, the delamination is modeled by subdividing the de-laminated region into separate waveguides and full damage model is established by imposing the kinematics. These models help to simulate wave propagation in such damaged beams to study the effect of damage on the wave response. Noise and vibration are often transmitted from the source to the other parts of the structure in the form of wave propagation. Thus, control of such wave transmission is essential for reduction of noise and vibration, which are the main cause of discomfort and in many cases cause failure of structure. Here, techniques for both passive and active controls of wave are proposed. For active control, a closed loop system is modeled using WSFE with magnetostrictive actuator for control of axial and flexural wave propagations in connected isotropic 1-D waveguides. The feedback is negative velocity and/or acceleration measured at different sensor points. A very new application of CNT reinforced composite for passive control of vibration and wave response is explored in this thesis. For this, a novel concept of nano-composite inserts is proposed. This insert can be made from CNTs dispersed in polymer. The high stiffness of the inserts helps to regulate the power flow in the form of wave propagation from the point of application of the loads to other parts of the structures. The length of the insert, volume fraction of CNTs and position are changed to achieve the required reduction in wave amplitudes. The entire thesis is split up into eight chapters. Chapter 1 presents a brief introduction, the motivation and objective of the thesis. Chapters 2 and 3 give a detail account of wavelet spectral finite element formulation for 1-D and 2-D isotropic waveguides, while Chapter 4 gives the same for composite waveguides. Chapter 5 brings out essential wave characteristics in carbon nanotubes and nano-composite structures, while Chapters 6 and 7 exclusively deal with application of WSFE to some real world problems. The thesis ends with summary and directions of future research. In summary, the thesis has brought out several new aspects of wave propagation in isotropic, composite and nano-composite structures. In addition to establishing wavelet spectral finite element as a useful tool for wave propagation analysis, several new techniques are presented, several new algorithm are proposed and several new concepts are explored.

Higher-order spectral analysis of acoustical waveforms can provide phase information that is not retained in calculations of power spectral density. In the propagation of high intensity sound, nonlinearity can cause substantial changes in the waveform as frequency components interact with one another. The bispectrum, which is one order higher than power spectral density, may provide a useful measure of nonlinearity in propagation by highlighting spectral regions of interaction. This thesis provides a review of the bispectrum, places it in the context of nonlinear acoustic propagation, and presents spectra calculated as a function of distance for numerically propagated acoustic waveforms. The calculated spectra include power spectral density, quad-spectral density, bispectrum, spatial derivative of the bispectrum, bicoherence, and skewness function. / text

In this paper a novel speaker recognition system is introduced. Automated speaker recognition has become increasingly popular to aid in crime investigations and authorization processes with the advances in computer science. Here, a recurrent neural network approach is used to learn to identify ten speakers within a set of 21 audio books. Audio signals are processed via spectral analysis into Mel Frequency Cepstral Coefficients that serve as speaker specific features, which are input to the neural network. The Long Short-Term Memory algorithm is examined for the first time within this area, with interesting results. Experiments are made as to find the optimum network model for the problem. These show that the network learns to identify the speakers well, text-independently, when the recording situation is the same. However the system has problems to recognize speakers from different recordings, which is probably due to noise sensitivity of the speech processing algorithm in use.

In this dissertation we will address two types of homogenization problems. The first one is a spectral problem in the realm of lower dimensional theories, whose physical motivation is the study of waves propagation in a domain of very small thickness and where it is introduced a very thin net of heterogeneities. Precisely, we consider an elliptic operator with "ε-periodic coefficients and the corresponding Dirichlet spectral problem in a three-dimensional bounded domain of small thickness δ. We study the asymptotic behavior of the spectrum as ε and δ tend to zero. This asymptotic behavior depends crucially on whether ε and δ are of the same order (δ ≈ ε), or ε is of order smaller than that of δ (δ = ετ , τ < 1), or ε is of order greater than that of δ (δ = ετ , τ > 1). We consider all three cases. The second problem concerns the study of multiscale homogenization problems with linear growth, aimed at the identification of effective energies for composite materials in the presence of fracture or cracks. Precisely, we characterize (n+1)-scale limit pairs (u,U) of sequences {(uεLN⌊Ω,Duε⌊Ω)}ε>0 ⊂ M(Ω;ℝd) × M(Ω;ℝd×N) whenever {uε}ε>0 is a bounded sequence in BV (Ω;ℝd). Using this characterization, we study the asymptotic behavior of periodically oscillating functionals with linear growth, defined in the space BV of functions of bounded variation and described by n ∈ ℕ microscales

In this thesis, the restorative effects of exposure to nature are examined through the lens of existing restoration theories. Limitations of existing theories, such as Attention Restoration Theory and Psycho-evolutionary Restoration Theory, are highlighted. To address the limitations of existing theories, an expanded theoretical framework is proposed: The expanded framework introduces a newly proposed neural mechanism and theory of restoration that build on existing theories by proposing a link to recently discovered reward systems in the ventral visual pathway. Results from six experiments provide consistent evidence to suggest that positive and negative responses to visual scenes are related to the low-level visuospatial properties of the scenes. Specifically, a discovery is made to suggest that the power of a limited visual spatial frequency range can consistently predict responses to natural, urban, and abstract scenes on measures of restoration (blink-rates, number of fixations, self-reported stress and pleasantness). This provides the first evidence to suggest that low-level visual properties of scenes may play an important role in affective and physiological responses to scenes. Furthermore, this newly discovered relationship provides a new way to objectively predict the relative restorative value of any given scene.

Global economic trends suggest that there is a need to generate sustainable renewable energy to meet growing global energy demands. Solar energy harnessed by concentrated photovoltaic (CPV) systems has a potential for strong contributions to future energy supplies. However, as a relatively new technology, there is still a need for considerable research into the relationship between the technology and the solar resource. Research into CPV systems was carried out at the University of Ottawa’s Solar Cells and Nanostructured Device Laboratory (SUNLAB), focusing on the acquisition and assessment of meteorological and local solar resource datasets as inputs to more complex system (cell) models for energy yield assessment. An algorithm aimed at estimating the spectral profile of direct normal irradiance (DNI) was created. The algorithm was designed to use easily sourced low resolution meteorological datasets, temporal band pass filter measurement and an atmospheric radiative transfer model to determine a location specific solar spectrum. Its core design involved the use of an optical depth parameterization algorithm based on a published objective regression algorithm. Initial results showed a spectral agreement that corresponds to 0.56% photo-current difference in a modeled CPV cell when compared to measured spectrum. The common procedures and datasets used for long term CPV energy yield assessment was investigated. The aim was to quantitatively de-convolute various factors, especially meteorological factors responsible for error bias in CPV energy yield evaluation. Over the time period from June 2011 to August 2012, the analysis found that neglecting spectral variations resulted in a ~2% overestimation of energy yields. It was shown that clouds have the dominant impact on CPV energy yields, at the 60% level.

In this thesis, we included results from two studies. The first one considered the effects of the blood volume changes, during blood donation, on the heart rate variability (HRV) measured, non-invasively, form electrocardiographic (ECG) and photoplethysmographic (PPG) signals. Our results showed that, during blood donation, there were no significant changes in the pulsatile area of PPG signal, while heart rate increased. No significant changes were noticed in HRV extracted from both signals. Error analysis between the HRV extracted from ECG and peak interval variability (PIV) suggested that the error during blood donation was increased which means that the use of PIV extracted from PPG signal, used as a replacement diagnostic tool in clinical applications, needs further investigations and should be carefully studied in non-stationary cardiovascular situations such as blood donation. The imbalance between the two branches of the autonomic nervous system, sympathetic and parasympathetic, vagal, may result in a harmful activation of myocardial tissues which cause arrhythmias and sudden cardiac death. Although the study of the sympathovagal balance have been attracting many researchers, further studies are needed to elucidate the effects of many kinds of drugs on the autonomic modulation of the cardiac muscle, specifically, the cells of sinoatrial (SA) node. The aim of the second part of this thesis was to assess the effects of calcium channel blocker (Verapamil), calcium channel sensitizer (Levosimendan), calcium chloride (CaCl2), the combinations of verapamil/ CaCl2, levosimendan/ CaCl2, and noradrenaline infusion on beat-to-beat cardiovascular variability represented, in this research, by systolic blood pressure variability (SBPV), and stroke volume variability (SVV) signals. We used Fat Fourier Transform (FFT) to evaluate the power spectral density of the fluctuations in both signals to evaluate the effects of short-term treatments with those drugs on the sympathovagal balance in normal rats. Then, we compared the spectra obtained from SBPV and SVV to decide which of these fluctuations along with corresponding spectrum was more able to provide a clear feedback about the autonomic nervous system. Our data suggests that there were a significant correlations between low- (LF), mid- (MF), and high-frequency (HF) spectra obtained from SBPV and SVV except between the HF spectra estimated from after the infusion of levosimendan where a poor correlation (r = 0.530, p = 0.281) was noticed. This that both HF components obtained provide different information regarding the autonomic nervous system modulation of the SA node cells, while the results obtained from the rest of experiments showed that both signals provide same information about the modulation of sympathetic and parasympathetic tone due to all stages of different drugs infusion studied in this thesis. Besides that, we found that both spectra may be used to track the fluctuations in the cardiac output as a result of the drugs infusion.

Ces dernières années ont vu l'émergence de la capture des modèles spatio-temporels (modélisation 4D) à partir d'images réelles, avec de nombreuses applications dans les domaines de post-production pour le cinéma, la science des sports, les études sociales, le divertissement, l'industrie de la publicité. A partir de plusieurs séquences vidéos, enregistrées à partir de points de vue variés, la modélisation 4D à partir de vidéos utilise des modèles spatio-temporels pour extraire des informations sur la géométrie et l'apparence de scènes réelles, permettant de les enregistrer et de les reproduire. Cette thèse traite du problème de la modélisation d'apparence.La disponibilité des donnée d'images offre de grands potentiels pour les reconstructions haute fidélité, mais nécessite des méthodes plus élaborées. En outre, les applications du monde réel nécessitent des rendus rapides et des flux réduits de données. Mais l'obtention de représentations d'apparence compactes, indépendantes du point de vue, et à grande résolution est toujours un problème ouvert.Pour obtenir ces caractéristiques, nous exprimons l'information visuelle de l'objet capturé dans un espace de texture commun. Les observations multi-caméra sont considérées comme des réalisations de l'apparence commune et un modèle linéaire est introduit pour matérialiser cette relation. Le modèle linéaire d'apparence proposé permet une première étude du problème de l'estimation d'apparence dans le cas multi-vue et expose les sources variées de bruit et les limitations intrinsèques du modèle.Basé sur ces observations, et afin d'exploiter l'information visuelle de la manière la plus efficace, nous améliorons la méthode en y intégrant un modèle de super-résolution 2D. Le modèle simule le procédé de capture d'image avec une concaténation d'opérations linéaires, générant les observation d'image des différents points de vue et permettant d'exploiter la redondance. Le problème de super-résolution multi-vue résultant est résolu par inférence bayésienne et une représentation haute-résolution d'apparence est fournie permettant de reproduire la texture de l'objet capturé avec grand détail.La composante temporelle est intégrée par la suite au modèle pour permettre d'y recouper l'information visuelle commune sous-jacente. En considérant des petits intervalles de temps ou l'apparence de l'objet ne change pas drastiquement, une représentation super-résolue cohérente temporellement est introduite. Elle explique l'ensemble des images de l'objet capturé dans cet intervalle. Grâce à l'inférence statistique Bayésienne, l'apparence construite permet des rendus avec une grande précision à partir de point de vue nouveau et à des instants différent dans l'intervalle de temps prédéfini.Pour améliorer l'estimation d'apparence d'avantage, l'inter-dépendance de la géométrie et de la photométrie est étudiée et exploitée. Les modélisations de la géométrie et de l'apparence sont unifiées dans le framework de super-résolution permettant une amélioration géométrique globale, ce qui donne à son tour une amélioration importante de l'apparence.Finalement pour encoder la variabilité de l'apparence dynamique des objets subissant plusieurs mouvements, une représentation indépendante du point de vue s'appuyant sur l'analyse en composantes principales est introduite. Cette représentation décompose la variabilité sous-jacente d'apparence en texture propres et déformations propres. La méthode proposée permet de reproduire les apparences de manière précise avec des représentation compactes. Il permet également l'interpolation et la complétion des apparences.Cette étude montre que la représentation compacte, indépendante du point de vue, et super-résolue proposée permet de confronter les nouvelles réalités du problème de modélisation d'apparence. Elle représente un contribution vers des représentations d'apparence 4D haute-qualité et ouvre de nouvelles directions de recherche dans ce domaine. / Capturing spatio-temporal models (4D modelling) from real world imagery has received a growing interest during the last years urged by the increasing demands of real-world applications and the tremendous amount of easily accessible image data. The general objective is to produce realistic representations of the world from captured video sequences. Although geometric modelling has already reached a high level of maturity, the appearance aspect has not been fully explored. The current thesis addresses the problem of appearance modelling for realistic spatio-temporal representations. We propose a view-independent, high resolution appearance representation that successfully encodes the high visual variability of objects under various movements.First, we introduce a common appearance space to express all the available visual information from the captured images. In this space we define the representation of the global appearance of the subject. We then introduce a linear image formation model to simulate the capturing process and to express the multi-camera observations as different realizations of the common appearance. Identifying that the principle of Super-Resolution technique governs also our multi-view scenario, we extend the image generative model to accommodate it. In our work, we use Bayesian inference to solve for the super-resolved common appearance.Second, we propose a temporally coherent appearance representation. We extend the image formation model to generateimages of the subject captured in a small time interval. Our starting point is the observation thatthe appearance of the subject does not change dramatically in a predefined small time interval and the visual information from each view and each frame corresponds to the same appearance representation.We use Bayesian inference to exploit the visual redundant as well as the hidden non-redundant information across time, in order to obtain an appearance representation with fine details.Third, we leverage the interdependency of geometry and photometry and use it toestimate appearance and geometry in a joint manner. We show that by jointly estimating both, we are able to enhance the geometry globally that in turn leads to a significant appearance improvement.Finally, to further encode the dynamic appearance variability of objects that undergo several movements, we cast the appearance modelling as a dimensionality reduction problem. We propose a view-independent representation which builds on PCA and decomposesthe underlying appearance variability into Eigen textures and Eigen warps. The proposed framework is shown to accurately reproduce appearances with compact representations and to resolve appearance interpolation and completion tasks.

La surveillance de phénomènes physiques à l’aide d’un réseau de capteurs (autonomes mais communicants) est fortement contrainte en consommation énergétique, principalement pour la transmission de données. Dans ce cadre, cette thèse propose des méthodes de traitement du signal permettant de réduire les communications sans compromettre la précision des calculs ultérieurs. La complexité de ces méthodes est réduite, de façon à ne consommer que peu d’énergie supplémentaire. Deux éléments servent à leur synthèse : la compression dès l’acquisition (Acquisition compressive) et la quantification grossière (sur 1 bit). D’abord, on étudie le corrélateur compressé, un estimateur qui permet d’évaluer les fonctions de corrélation, temps de retard et densités spectrales en exploitant directement des signaux compressés. Ses performances sont comparées au corrélateur usuel. Si le signal à traiter possède un support spectral étroit, l’estimateur proposé s’avère sensiblement meilleur que l’usuel. Ensuite, inspirés par les corrélateurs à forte quantification des années 50 et 60, deux nouveaux corrélateurs sont étudiés : le compressé sur 1 bit et le compressé hybride, qui peuvent également surpasser les performances de leurs contreparties non-compressées. Finalement, on montre la pertinence de ces méthodes pour les applications envisagées à travers l’exploitation de données réelles. / Monitoring physical phenomena by using a network of sensors (autonomous but interconnected) is highly constrained in energy consumption, mainly for data transmission. In this context, this thesis proposes signal processing tools to reduce communications without compromising computational accuracy in subsequent calculations. The complexity of these methods is reduced, so as to consume only little additional energy. Our two building blocks are compression during signal acquisition (Compressive Sensing) and CoarseQuantization (1 bit). We first study the Compressed Correlator, an estimator which allows for evaluating correlation functions, time-delay, and spectral densities directly from compressed signals. Its performance is compared with the usual correlator. As we show, if the signal of interest has limited frequency content, the proposed estimator significantly outperforms theconventional correlator. Then, inspired by the coarse quantization correlators from the 50s and 60s, two new correlators are studied: The 1-bit Compressed and the Hybrid Compressed, which can also outperform their uncompressed counterparts. Finally, we show the applicability of these methods in the context of interest through the exploitation of real data.

This thesis represents an in-depth empirical study of the dependence structures within the term structure of interest rates. Firstly, a comprehensive overview of term structure modelling literature and methods is provided together with a summary of theoretical notions regarding the use of high-frequency data and spectral analysis. Contrary to most studies, the frequency-domain approach is employed, with a special focus on dependency across various quantiles of the joint distribution of the term structure. The main results are obtained using the quantile cross-spectral analysis, a new robust and non-parametric method allowing to uncover dependence structures in quantiles of the joint distribution of multivariate time series. The results are estimated using a dataset consisting of 15 years worth of high-frequency tick-by-tick time series of US Treasury futures. Complex dependence structures are revealed showing signs of both cyclicity and dependence in various parts of the joint distribution of the term structure in the frequency domain. JEL Classification C49, C55, C58, E43, G12, G13 Keywords term structure of interest rates, yield curves, high-frequency analysis, spectral analysis, inter- est rate futures Author's e-mail adam.nedved@fsv.cuni.cz Supervisor's e-mail barunik@fsv.cuni.cz