Les travaux présentés proposent une nouvelle approche du contrôle actif du bruit à l'intérieur des hélicoptères. Les stratégies de contrôle s'attachaient à réduire le bruit moyennes fréquences en grande partie produit par la boite de transmission principale. Nous proposons d'utiliser le contrôle actif pour traiter les basses fréquences dans la cabine. Les niveaux mis en jeu dans ces gammes de fréquences sont tels qu'un contrôle acoustique nécessite une attention particulière dans la confection des sources antibruits.<br />Nous avons montré l'intérêt de scinder le problème en deux parties. Un premier dispositif est conçu pour traiter la fréquence produite par le rotor principal de l'autogire, et un second dispositif est destiné à traiter la partie large bande du bruit basses fréquences. Nous proposons une méthode spécifique pour la conception des sources antibruits.<br />Compte tenu de la faible disponibilité des aéronefs pour les essais du système, nous avons eu recours à une maquette à l'échelle 1. Nous avons proposé une méthode originale permettant d'y produire un champ primaire perceptivement similaire à celui présent sur l'hélicoptère.<br />Nous avons validé notre stratégie de contrôle par des simulations d'algorithmes FXLMS et IMC-FXLMS. Ces simulations, basées sur des mesures issues d'une campagne d'essaisl, ont pu être validées par des essais du dispositif sur la maquette mettant en œuvre un contrôleur et les sources que nous avons développé. Au cours de l'étude, nous avons participé à la mise au point du matériel embarquable (contrôleur, ampli. et sources). Ce matériel nous a permis de réaliser des essais en vol en fin de thèse qui ont montré l'efficacité du dispositif.

This thesis strives to enhance the performance of delta-sigma modulators in two areas: increasing their speed and enabling their operation in a low voltage environment. Parallelism based on time-interleaving can be used to increase the speed of delta-sigma modulators. A novel single-path time-interleaved architecture is derived and analyzed. Finite opamp gain and bandwidth result in a mismatch between the noise transfer functions of the internal quantizers which degrades the performance of the new modulator. Two techniques are presented to mitigate the mismatch problem: a hybrid topology where the first stage uses multiple integrators while the rest of the modulator uses a single path of integrators and a digital calibration method. The input-feedforward technique removes the input-signal component from the internal nodes of delta-sigma modulators. The removal of the signal component reduces the signal swing and distortion requirements for the opamps. These characteristics enable the reliable implementation of delta-sigma modulators in modern CMOS technology. Two implementation issues for modulators with input-feedforward are considered. First, the drawback of the analog adder at the quantizer input is identified and the capacitive input feedforward technique is introduced to eliminate the adder. Second, the double sampled input technique is proposed to remove the critical path generate by the input feedforward path. Novel input-feedforward delta-sigma architecture is proposed. The new digital input feedforward (DIFF) modulator maintains the low swing and low distortion requirements of the input feedforward technique, it eliminates the analog adder at the quantizer input, and it improves the achievable resolution. To demonstrate these advantages, a configurable delta-sigma modulator which can operate as a feedback topology or in DIFF mode is implemented in 0.18μm CMOS technology. Both modulators operate at 20MHz clock with an oversampling ratio of 8. The power consumption in the DIFF mode is 22mW and in feedback mode is 19mW. However, the DIFF mode achieves a peak SNDR of 73.7dB (77.1dB peak SNR) while the feedback mode achieves a peak SNDR of 64.3dB (65.9dB peak SNR). Therefore, the energy required per conversion step for the DIFF architecture (2.2 pJ/step) is less than half of that required by the feedback architecture (5.7 pJ/step).

Wireless communication systems such as multipoint communication systems (MCS) are becoming attractive as cost-effective means for providing network access in sparsely populated, rugged, or developing areas of the world. Since the radio spectrum is limited, it is desirable to use spectrally efficient modulation methods such as quadrature amplitude modulation (QAM) for high data rate channels. Many MCS employ time division multiple access (TDMA) and/or time division duplexing (TDD) techniques, in which transmissions operate in bursts. In many cases, a preamble of known symbols is appended to the beginning of each burst for carrier and symbol timing recovery (symbol timing is assumed known in this thesis). Preamble symbols consume bandwidth and power and are not used to convey information. In order for burst-mode communications to provide efficient data throughput, the synchronization time must be short compared to the user data portion of the burst. <p> Traditional methods of communication system synchronization such as phase-locked loops (PLLs) have demonstrated reduced performance when operated in burst-mode systems. In this thesis, a feedforward (FF) digital carrier recovery technique to achieve rapid carrier synchronization is proposed. The estimation algorithms for determining carrier offsets in carrier acquisition and tracking in a linear channel environment corrupted by additive white Gaussian noise (AWGN) are described. The estimation algorithms are derived based on the theory of maximum likelihood (ML) parameter estimation. The estimations include data-aided (DA) carrier frequency and phase estimations in acquisition and non-data-aided (NDA) carrier phase estimation in tracking. The DA carrier frequency and phase estimation algorithms are based on oversampling of a known preamble. The NDA carrier phase estimation makes use of symbol timing knowledge and estimates are extracted from the random data portion of the burst. The algorithms have been simulated and tested using Matlab® to verify their functionalities. The performance of these estimators is also evaluated in the burst-mode operations for 16-QAM and compared in the presence of non-ideal conditions (frequency offset, phase offset, and AWGN). The simulation results show that the carrier recovery techniques presented in this thesis proved to be applicable to the modulation schemes of 16-QAM. The simulations demonstrate that the techniques provide a fast carrier acquisition using a short preamble (about 111 symbols) and are suitable for burst-mode communication systems.

Wireless communication systems such as multipoint communication systems (MCS) are becoming attractive as cost-effective means for providing network access in sparsely populated, rugged, or developing areas of the world. Since the radio spectrum is limited, it is desirable to use spectrally efficient modulation methods such as quadrature amplitude modulation (QAM) for high data rate channels. Many MCS employ time division multiple access (TDMA) and/or time division duplexing (TDD) techniques, in which transmissions operate in bursts. In many cases, a preamble of known symbols is appended to the beginning of each burst for carrier and symbol timing recovery (symbol timing is assumed known in this thesis). Preamble symbols consume bandwidth and power and are not used to convey information. In order for burst-mode communications to provide efficient data throughput, the synchronization time must be short compared to the user data portion of the burst. <p> Traditional methods of communication system synchronization such as phase-locked loops (PLLs) have demonstrated reduced performance when operated in burst-mode systems. In this thesis, a feedforward (FF) digital carrier recovery technique to achieve rapid carrier synchronization is proposed. The estimation algorithms for determining carrier offsets in carrier acquisition and tracking in a linear channel environment corrupted by additive white Gaussian noise (AWGN) are described. The estimation algorithms are derived based on the theory of maximum likelihood (ML) parameter estimation. The estimations include data-aided (DA) carrier frequency and phase estimations in acquisition and non-data-aided (NDA) carrier phase estimation in tracking. The DA carrier frequency and phase estimation algorithms are based on oversampling of a known preamble. The NDA carrier phase estimation makes use of symbol timing knowledge and estimates are extracted from the random data portion of the burst. The algorithms have been simulated and tested using Matlab® to verify their functionalities. The performance of these estimators is also evaluated in the burst-mode operations for 16-QAM and compared in the presence of non-ideal conditions (frequency offset, phase offset, and AWGN). The simulation results show that the carrier recovery techniques presented in this thesis proved to be applicable to the modulation schemes of 16-QAM. The simulations demonstrate that the techniques provide a fast carrier acquisition using a short preamble (about 111 symbols) and are suitable for burst-mode communication systems.

This thesis strives to enhance the performance of delta-sigma modulators in two areas: increasing their speed and enabling their operation in a low voltage environment. Parallelism based on time-interleaving can be used to increase the speed of delta-sigma modulators. A novel single-path time-interleaved architecture is derived and analyzed. Finite opamp gain and bandwidth result in a mismatch between the noise transfer functions of the internal quantizers which degrades the performance of the new modulator. Two techniques are presented to mitigate the mismatch problem: a hybrid topology where the first stage uses multiple integrators while the rest of the modulator uses a single path of integrators and a digital calibration method. The input-feedforward technique removes the input-signal component from the internal nodes of delta-sigma modulators. The removal of the signal component reduces the signal swing and distortion requirements for the opamps. These characteristics enable the reliable implementation of delta-sigma modulators in modern CMOS technology. Two implementation issues for modulators with input-feedforward are considered. First, the drawback of the analog adder at the quantizer input is identified and the capacitive input feedforward technique is introduced to eliminate the adder. Second, the double sampled input technique is proposed to remove the critical path generate by the input feedforward path. Novel input-feedforward delta-sigma architecture is proposed. The new digital input feedforward (DIFF) modulator maintains the low swing and low distortion requirements of the input feedforward technique, it eliminates the analog adder at the quantizer input, and it improves the achievable resolution. To demonstrate these advantages, a configurable delta-sigma modulator which can operate as a feedback topology or in DIFF mode is implemented in 0.18μm CMOS technology. Both modulators operate at 20MHz clock with an oversampling ratio of 8. The power consumption in the DIFF mode is 22mW and in feedback mode is 19mW. However, the DIFF mode achieves a peak SNDR of 73.7dB (77.1dB peak SNR) while the feedback mode achieves a peak SNDR of 64.3dB (65.9dB peak SNR). Therefore, the energy required per conversion step for the DIFF architecture (2.2 pJ/step) is less than half of that required by the feedback architecture (5.7 pJ/step).

El objetivo de esta Tesis Doctoral ha sido básicamente el de realizar un estudio comparativo sobre los diferentes métodos existentes para resolver diversos aspectos referentes al diseño de la arquitectura de red neuronal Multilayer Feedforward, en problemas de clasificación de redes neuronales.Los aspectos de diseño de la arquitectura de red neuronal estudiados han sido: codificación de entradas desconocidas, selección de la información de entrada a la red, selección del número de unidades ocultas, influencia en la capacidad de generalización del número de capas ocultas e inicialización de pesos de la red.Para cada uno de los aspectos se ha realizado un estudio comparativo de los diferentes métodos existentes para resolver dicho problema. Como resultado recomendamos finalmente el uso de los mejores métodos a la hora de realizar una aplicación concreta. / The objective of this Doctoral Thesis was to carry a comparative study on several existent methods in order to solve different aspects of the design of Multilayer Feedforward architecture, in neural networks classification problems.The aspects of design studied were: handling unknown input information, input selection, selection of the number of hidden units, influence in the generalization capability of the number of hidden layers and weight initialization.For each one of these aspects, we carried out a comparative study of several existent methods in order to solve the problem. We recommend the use of the best methods in order to develop a concrete application.

Critical steps in the design process is the gathering of data, processing the data into a useful form of information (a design concept) which meets specific needs, passing this refined design solution down the path to production, where it is released into the larger environment. With in the designing process there are multiple feedback loops as the solution becomes more refined. Even as it reaches the final end user, other design refinement feedback loops continue as new and improved products or systems. Along with the interdisciplinary teams involved with the product/system development, the more complexity the product or system becomes the more critical the organization of the data becomes. This paper will present and test a concept of a design feedback and feed forward communication tool for product/system design that uses Dr. Walter A. Schaer s Three Functions of an Artifact as the methodological structure for design development. The essence of this design tool is the merging of a new communication system within an existing methodology of organizing complex systems into a morphological matrix, developed by Dr. Walter A. Schaer, based on the Charles Morris s work on semiotics. This communication tool is a new feedback / feed forward mechanism which correspond with the semiotic structure in a morphological matrix to assist the designer develop design solutions. The research will measure the success rate of the tool in the design process, examine of how the designers took advantage of the new tool, and evaluate their perception of its usefulness.

This paper presents a distributed feedback and feedforward of discrete-time delta sigma modulator applications in the radio. We know the delta-sigma modulator using oversampling and noise shaping technique, thus we can relax the specifications of the components. This paper described the architectural differences and compare, the in-band signal is less sensitive to noise interference, and improve the resolution of the circuit. In the resonator, a simple structure with a small number of capacitor in resonator circuit. This paper uses the TSMC 0.18£gm process parameters to the simulation, implementation, and measurement. Our fourth-order discrete-time delta-sigma modulator specifications as follows: the input signal frequency is 10.7MHz, the sampling frequency is 42.8MHz, the signal bandwidth is 200kHz, oversampling rate is 107, and one bit quantizer.

<p>Arbetet som presenteras i den här rapporten handlar om ansiktsigenkänning från webbkameror med hjälp av principal component analysis samt artificiella neurala nätverk av typen feedforward. Arbetet förbättrar tekniken med hjälp av filterbaserade metoder som bland annat används inom ansiktsdetektering. Dessa filter bygger på att skicka med redundant data av delregioner av ansiktet.</p>

Dynamic security limits the power transfer capacity between regions and therefore has an economic impact. The power modulation control of high-voltage direct current (HVDC) links can improve the dynamic security of the power system. Having several HVDC links in a system creates the opportunity to coordinate such control, and coordination also ensures that negative interactions do not occur among the controllable devices. This thesis aims to increase dynamic security by coordinating HVDC links, as an alternative to decreasing the transfer capacity. This thesis contributes four control approaches for increasing the dynamic stability, based on feedforward control, adaptive control, optimal control, and exact-feedback linearization control. Depending on the available measurements, dynamic system model, and system topology, one of the developed methods can be applied. The wide-area measurement system provides the central controller with real-time data and sends control signals to the HVDC links. The feedforward controller applies rapid power dispatch, and the strategy used here is to link the N-1 criterion between two systems. The adaptive controller uses the modal analysis approach; based on forecasted load paths, the controller gains are adaptively adjusted to maximize the damping in the system. The optimal controller is designed based on an estimated reduced-order model; system identification develops the model based on the system response. The exact-feedback linearization approach uses a pre-feedback loop to cancel the nonlinearities; a stabilizing controller is designed for the remaining linear system. The conclusion is that coordinating the HVDC links improves the dynamic stability, which makes it possible to increase the transfer capacity. This conclusion is also supported by simulations of each control approach. / QC 20110302