Le présent travail traite de problèmes particuliers en traitement de signal soulevés par l'holographie digitale. Le premier chapitre présente l'holographie et l'holographie digitale. Le deuxième chapitre aborde le problème des artéfacts apparaissant sur le bord des hologrammes lorsque ceux-ci sont propagés numériquement. Le troisième chapitre propose une méthode automatique pour éliminer la phase de fond d'un hologramme. Le quatrième chapitre montre comment améliorer la visibilité des hologrammes en supprimant les objets diffractants. Le cinquième chapitre expose un logiciel d'aide à l'analyse des hologrammes enregistrés dans le temps. On montre la capacité du logiciel à effectuer du tracking d'objets en trois dimensions à posteriori. Le sixième chapitre traite de la reconnaissance de forme dans les hologrammes, le problème de la reconnaissance de particules micrométriques est abordé. Des conclusions et perspectives terminent le travail. / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

No description available.

Depuis plusieurs années, les technologies de l’information et de la communication connaissent un développement spectaculaire. Le trafic d’information augmente exponentiellement alors que l’évolution de l’efficacité énergétique des réseaux, définie généralement comme le rapport entre l’énergie utile en sortie du système et l’énergie totale fournie au système, n’augmente que de manière linéaire. Pour soutenir la demande de trafic, il est alors nécessaire de multiplier les équipements réseaux ce qui augment la facture énergétique des opérateurs et les émissions de gaz carbonique. Devant l’urgence de la situation, de nombreux projets et consortiums ont été créés avec l’objectif de renforcer l’efficacité énergétique des réseaux. Par exemple, le consortium GreenTouch™ fondé en 2010 a pour objectif d’améliorer l’efficacité énergétique des réseaux d’un facteur 1000 d’ici 2015 et de réduire leur consommation énergétique de 90% d’ici 2020 (par rapport à 2010). Le projet METIS a pour objectif de définir les télécommunications de 2020 (5G) cherche des solutions efficaces minimisant les coûts, les ressources et la consommation énergétique. Le projet MiWEBA qui intègre les ondes millimétrique aux communications mobiles dans le but d’augmenter la capacité du réseau d’un facteur 1000 avec un coût raisonnable.Dans cette thèse, nous abordons deux solutions répondant aux challenges de la 5G et permettant de réduire les niveaux des puissances rayonnées par les systèmes de télécommunications : l’amélioration de la couche physique via les techniques de focalisation spatiale et l’adaptation de lien multi-technologies. L’objectif principal de la thèse est de réaliser des transmissions de données très haut débit à forte efficacité énergétique à l’aide d’une connectivité multi-technologies sans coutures y compris dans des réseaux comprenant une multitude de périphériques à faible complexité.L’organisation de la thèse est la suivante : tout d’abord, l’implémentation de techniques de focalisation spatiale, le Retournement Temporel (RT) et l’EGT (Equal Gain Transmission), dans un système utilisant les dernières évolutions du standard IEEE 802.11, notamment l’intégration d’une composante 60 GHz, est étudiée. Les performances de ces techniques sont analysées de manière à identifier les conditions optimales de leur utilisation et il est montré qu’avec un minimum de deux antennes de transmission, elles permettent d’améliorer les performances énergétiques des systèmes tout en ayant une faible complexité d’implémentation. Des modifications du standard permettant d’augmenter l’efficacité énergétique sont également proposées. Une analyse cross-layer est ensuite réalisée afin de définir protocoles d’estimation de canal tenant compte des spécificités du RT. De plus, les améliorations énergétiques liées à l’utilisation du RT sont confirmées en tenant compte des dégradations des performances de la couche de liaison. Enfin, une métrique d’occupation temporelle est définie et associée à des techniques d’adaptation de lien mono et multi-technologies existantes dans le but d’améliorer la sélection d‘interfaces de transmission pour augmenter l’efficacité énergétique sans affecter les performances des systèmes utilisant un accès aléatoire au canal de communication. L’utilisation de cette métrique permet, dans certaines configurations, d’augmenter d’un facteur sept l’efficacité énergétique d’un système utilisant le Wi-Fi et, dans le cas multi-technologies, de répartir le trafic sur les différentes interfaces de transmission afin d’éviter la saturation des canaux de communication. / Over the past few years, information and communications technologies have experienced a spectacular development. Traffic demand grows exponentially while energy efficiency of networks –usually defined as the ratio between delivered power and supplied power- only increases linearly. In order to supply the data traffic, it is necessary to duplicate networks equipments increasing energy costs for operators and carbon dioxide emissions. To face this dramatic situation, several projects and consortiums have been created and aim to increase networks energy efficiency. For instance, the GreenTouch™ consortium funded in 2010 aims to improve networks energy efficiency by a factor 1000 by 2015 and to reduce energy consumption of TICs by 90% by 2020 (compared with 2010 levels). METIS project aims to define 2020s communications (5G) search for efficient solutions with low costs, low resources use and low power consumption of systems. MiWEBA project brings millimeter-waves into the mobile radio word to extend the network capacity by 1000 at reasonable cost.In this thesis we study two solutions to answer 5G challenges and reduce radiated power levels of telecommunications systems: physical layer improvements through spatial focalization techniques and multi-technologies link adaptation. This thesis main goal is to achieve very high data transmission rates with high energy efficiency and seamless connectivity using multiple-interfaces technologies even in networks composed of a high number of low-complexity devices.The thesis is organized as follows: first, Time Reversal (TR) and Equal Gain Transmission (EGT) implementation is studied and applied to a system using the latest amendments of IEEE 802.11 standard, including 60 GHz transmissions. Performance of these techniques is analyzed in order to identify optimal conditions of use and it is shown that they only require a minimum of two transmit antennas to increase the energetic performance of systems while having a low implementation complexity. Standard modifications allowing an increase of energy efficiency are proposed. Then, a cross-layer analysis is performed to define channel estimation protocols taking TR specificities into account. Moreover, TR energetic benefits are validated considering its drawbacks at the link layer level. Finally, a time-availability metric is defined and associated to existing mono or multi-technologies adaptation techniques and aim to increase energy efficiency without degrading the performance of systems using random access to the communication channel. This metrics allows –in some configurations- to increase by a factor seven the energy efficiency of a system using Wi-Fi and, in a multi-technologies scenario, to allocate the data traffic to the several interfaces in order to avoid communication channels saturation.

The purpose of this project is to improve the existing version of the Sequential Alignment and Position Verification System (SAPVS) for functional proton radiosurgery and to evaluate its performance after improvement .

One of the challenges of utilizing higher frequencies in the RF spectrum, for any number of applications, is the hardware constraints of analog-to-digital converters (ADCs). Since mid-20th century, we have accepted the Nyquist-Shannon Sampling Theorem in that we need to sample a signal at twice the max frequency component in order to reconstruct it. Compressive Sampling (CS) offers a possible solution of sampling sub-Nyquist and reconstructing using convex programming techniques. There has been significant advancements in CS research and development (more notably since 2004), but still nothing to the advantage of everyday use. Not for lack of theoretical use and mathematical proof, but because of no implementation work. There has been little work on hardware in finding the realistic constraints of a working CS system used for digital signal process (DSP). Any parameters used in a system is usually assumed based on stochastic models, but not optimized towards a specific application. This thesis aims to address a minimal viable platform to implement compressive sensing if applied to a wireless sensor network (WSN), as well as address certain parameters of CS theory to be modified depending on the application.

A novel scheme of optimizing the individual components of a phase locked loop (PLL) which is used for stable clock generation and synchronization of signals is considered in this work. Verilog-A is used for the high level system design of the main components of the PLL, followed by the individual component wise optimization. The design of experiments (DOE) approach to optimize the analog, 45nm voltage controlled oscillator (VCO) is presented. Also a mixed signal analysis using the analog and digital Verilog behavior of components is studied. Overall a high level system design of a PLL, a systematic optimization of each of its components, and an analog and mixed signal behavioral design approach have been implemented using cadence custom IC design tools.

Many communications signal formats are not power efficient because of their large peak-to-average power ratios (PARs). Moreover, in the presence of nonlinear devices such as power amplifiers (PAs) or mixers, the non-constant-modulus signals may generate both in-band distortion and out-of-band interference. Backing off the signal to the linear region of the device further reduces the system power efficiency. To improve the power efficiency of the communication system, one can pursue two approaches: i) linearize the PA; ii) reduce the high PAR of the input signal. In this dissertation, we first explore the optimal nonlinearity under the peak power constraint. We show that the optimal nonlinearity is a soft limiter with a specific gain calculated based on the peak power limit, noise variance, and the probability density function of the input amplitude. The result is also extended to the fading channel case. Next, we focus on digital baseband predistortion linearization for power amplifiers with memory effects. We build a high-speed wireless test-bed and carry out digital baseband predistortion linearization experiments. To implement adaptive PA linearization in wireless handsets, we propose an adaptive digital predistortion linearization architecture that utilizes existing components of the wireless transceiver to fulfill the adaptive predistorter training functionality. We then investigate the topic of PAR reduction for OFDM signals and forward link CDMA signals. To reduce the PAR of the OFDM signal, we propose a dynamic selected mapping (DSLM) algorithm with a two-buffer structure to reduce the computational requirement of the SLM method without sacrificing the PAR reduction capability. To reduce the PAR of the forward link CDMA signal, we propose a new PAR reduction algorithm by introducing a relative offset between the in-phase branch and the quadrature branch of the transmission system.

We investigated two design strategies that would allow us to efficiently process audio signals on embedded systems such as mobile phones and portable electronics. In the first strategy, we exploit properties of the human auditory system to process audio signals. We designed a sound enhancement algorithm to make piezoelectric loudspeakers sound "richer" and "fuller," using a combination of bass extension and dynamic range compression. We also developed an audio energy reduction algorithm for loudspeaker power management by suppressing signal energy below the masking threshold. In the second strategy, we use low-power analog circuits to process the signal before digitizing it. We designed an analog front-end for sound detection and implemented it on a field programmable analog array (FPAA). The sound classifier front-end can be used in a wide range of applications because programmable floating-gate transistors are employed to store classifier weights. Moreover, we incorporated a feature selection algorithm to simplify the analog front-end. A machine learning algorithm AdaBoost is used to select the most relevant features for a particular sound detection application. We also designed the circuits to implement the AdaBoost-based analog classifier.

It is well known that communication systems employing multiple transmit and multiple receive antennas provide high data rates along with increased reliability. It has been shown that coding across both spatial and temporal domains together, called Space-Time Coding (STC), achieves, a diversity order equal to the product of the number of transmit and receive antennas. Space-Time Block Codes (STBC) achieving the maximum diversity is called full-diversity STBCs. An STBC is called information-lossless, if the structure of it is such that the maximum mutual information of the resulting equivalent channel is equal to the capacity of the channel. This thesis deals with high-rate and information-lossless STBCs obtained from certain matrix algebras called Crossed-Product Algebras. First we give constructions of high-rate STBCs using both commutative and non-commutative matrix algebras obtained from appropriate representations of extensions of the field of rational numbers. In the case of commutative algebras, we restrict ourselves to fields and call the STBCs obtained from them as STBCs from field extensions. In the case of non-commutative algebras, we consider only the class of crossed-product algebras. For the case of field extensions, we first construct high-rate; full-diversity STBCs for arbitrary number of transmit antennas, over arbitrary apriori specified signal sets. Then we obtain a closed form expression for the coding gain of these STBCs and give a tight lower bound on the coding gain of some of these STBCs. This lower bound in certain cases indicates that some of the STBCs from field extensions are optimal m the sense of coding gain. We then show that the STBCs from field extensions are information-lossy. However, we also show that the finite-signal-set capacity of the STBCs from field extensions can be improved by increasing the symbol rate of the STBCs. The simulation results presented show that our high-rate STBCs perform better than the rate-1 STBCs in terms of the bit error rate performance. Then we proceed to present a construction of high-rate STBCs from crossed-product algebras. After giving a sufficient condition on the crossed-product algebras under which the resulting STBCs are information-lossless, we identify few classes of crossed-product algebras that satisfy this sufficient condition and also some classes of crossed-product algebras which are division algebras which lead to full-diversity STBCs. We present simulation results to show that the STBCs from crossed-product algebras perform better than the well-known codes m terms of the bit error rate. Finally, we introduce the notion of asymptotic-information-lossless (AILL) designs and give a necessary and sufficient condition under which a linear design is an AILL design. Analogous to the condition that a design has to be a full-rank design to achieve the point corresponding to the maximum diversity of the optimal diversity-multiplexing tradeoff, we show that a design has to be AILL to achieve the point corresponding to the maximum multiplexing gain of the optimal diversity-multiplexing tradeoff. Using the notion of AILL designs, we give a lower bound on the diversity-multiplexing tradeoff achieved by the STBCs from both field extensions and division algebras. The lower bound for STBCs obtained from division algebras indicates that they achieve the two extreme points, 1 e, zero multiplexing gain and zero diversity gain, of the optimal diversity-multiplexing tradeoff. Also, we show by simulation results that STBCs from division algebras achieves all the points on the optimal diversity-multiplexing tradeoff for n transmit and n receive antennas, where n = 2, 3, 4.

The demand for high-performance and power scalable DSP processors for telecommunication and portable devices has increased significantly in recent years. The Fast Fourier Transform (FFT) computation is essential to such designs. This work presents a switch-based architecture to design radix-2 FFT processors. The processor employs M processing elements, 2M memory arrays and M Read Only Memories (ROMs). One processing element performs one radix-2 butterfly operation. The memory arrays are designed as single-port memory, where each has a size of N/(2M); N is the number of FFT points. Compared with a single processing element, this approach provides a speedup of M. If not addressed, memory collisions degrade the processor performance. A novel algorithm to detect and resolve the collisions is presented. When a collision is detected, a memory management operation is executed. The performance of the switch architecture can be further enhanced by pipelining the design, where each pipeline stage employs a switch component. The result is a speedup of Mlog2N compared with a single processing element performance. The utilization of single-port memory reduces the design complexities and area. Furthermore, memory arrays significantly reduce power compared with the delay elements used in some FFT processors. The switch-based architecture facilitates deactivating processing elements for power scalability. It also facilitates implementing different FFT sizes. The VLSI implementation of a non-pipeline switch-based processor is presented. Matlab simulations are conducted to analyze the performance. The timing, power and area results from RTL, synthesis and layout simulations are discussed and compared with other processors. / text