La détection cohérente combinée avec le traitement du signal s’est imposée comme le standard pour les systèmes de communications optiques longue distance à 100 Gb/s (mono-porteuse) et au-delà. Avec l'avènement des convertisseurs numérique-analogique à haute vitesse et haute résolution, la génération de formats de modulation d'ordre supérieure avec filtrage numérique est devenue possible, favorisant l’émergence de transmissions à forte densité spectrale. En outre, la généralisation des liaisons non gérées en dispersion permet une modélisation analytique du canal optique et favorise l'utilisation d’outils puissants de la théorie de l'information et du traitement du signal. En se fondant sur ces outils, de nouveaux formats de modulation à entrelacement temporel dits hybrides et formats multidimensionnels sont étudiés et mise en oeuvre expérimentalement. Leur impact sur les algorithmes de traitement du signal et sur le débit d'information atteignable est analysé en détail. La conception de transpondeurs de prochaine génération à 400 Gb/ s et 1 Tb/s reposant sur des signaux à débit-symbole élevé est également étudiée. Ces systèmes sont intéressants pour réduire le coût par bit en augmentant la capacité émise par transpondeur. L'élaboration d'algorithmes de traitement du signal avancés associés à l’utilisation de composants optoélectroniques à l'état de l'art ont permis la démonstration d’expériences records: d’une part la première transmission mono-porteuse à 400 Gb/s sur une distance transatlantique (pour une efficacité spectrale de 6 b/s/Hz) d’autre part la première transmission à 1 Tb/s basée sur la synthèse en parallèle de plusieurs tranches spectrales (8 b/s/Hz) / Coherent detection in combination with digital signal processing is now the de facto standard for long-haul high capacity optical communications systems operating at 100 Gb/s per channel and beyond. With the advent of high-speed high-resolution digital-to-analog converters, generation of high order modulation formats with digital pulse shaping has become possible allowing the increase of system spectral efficiency. Furthermore, the widespread use of transmission links without in-line dispersion compensation enables elegant analytical optical channel modeling which facilitates the use of powerful tools from information theory and digital signal processing. Relying on these aforementioned tools, the introduction of time-interleaved hybrid modulation formats, multi-dimensional modulation formats, and alternative quadrature amplitude modulation formats is investigated in high-speed optical transmission systems. Their impact on signal processing algorithms and achievable information rate over optical links is studied in detail. Next, the design of next generation transponders based on high symbol rate signals operating at 400 Gb/s and 1 Tb/s is investigated. These systems are attractive to reduce the cost per bit as more capacity can be integrated in a single transponder. Thanks to the development of advanced signal processing algorithms combined with state-of-the-art opto-electronic components, record high-capacity transmission experiments are demonstrated: the first single carrier 400 Gb/s transmission over transatlantic distance (at 6 b/s/Hz) and the first 1 Tb/s net data rate transmission based on the parallel synthesis of multiple spectral slices (at 8 b/s/Hz)

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第20740号 / 情博第654号 / 新制||情||113(附属図書館) / 京都大学大学院情報学研究科通信情報システム専攻 / (主査)教授 守倉 正博, 教授 大木 英司, 教授 梅野 健 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

In recent years, the Brigham Young University (BYU) Radio Astronomy Systems group has developed phased-array feeds and the data acquisition processing systems necessary to perform radio astronomy observations. This thesis describes the development and testing of a real-time digital beamforming system that reduces both the time required to process phased-array feed data and the disk space used to record this data compared to post-processing beamforming systems. A real-data experiment is also discussed in this thesis, which focuses on some of the data post-processing required for one of BYU's data acquisition systems.Radio-frequency interference mitigation techniques for phased-array feed radio astronomy have been studied for several years, but the effect that these techniques have on weak-signal detection is not well understood. This thesis provides analysis of a simulated weak-source observation for the Green Bank 20-meter telescope and BYU 19 element phasedarray feed with radio-frequency interference present. Interference mitigation techniques are shown to reduce the detectability of weak sources compared with the no interference case, but it is also shown that a weak source can be detected that would otherwise be masked by interference.

Coherent detection employing multilevel modulation format has become one of the most promising technologies for next generation high speed transmission system due to the high power and spectral efficiencies. With the powerful digital signal processing (DSP), coherent optical receivers allow the significant equalization of chromatic dispersion (CD), polarization mode dispersion (PMD), phase noise (PN) and nonlinear effects in the electrical domain. Recently, the realizations of these DSP algorithms for mitigating the channel distortions in the transmission system are the most attractive investigations.  The CD equalization can be performed by the digital filters developed in the time and the frequency domain, which can suppress the fiber dispersion effectively. The PMD compensation is usually performed in the time domain with the adaptive least mean square (LMS) and constant modulus algorithms (CMA) equalization. Feed-forward and feed-back carrier phase estimation algorithms are employed to mitigate the phase noise from the transmitter and local oscillator lasers. The fiber nonlinearities are compensated by using the digital backward propagation methods based on solving the nolinear Schrodinger (NLS) equation and the Manakov equation.  In this dissertation, we present a comparative analysis of three digital filters for chromatic dispersion compensation, an analytical evaluation of carrier phase estimation with digital equalization enhanced phase noise and a brief discussion for PMD adaptive equalization. To implement these investigations, a 112-Gbit/s non-return-to-zero polarization division multiplexed quadrature phase shift keying (NRZ-PDM-QPSK) coherent transmission system is realized in the VPI simulation platform. With the coherent transmission system, these CD equalizers have been compared by evaluating their applicability for different fiber lengths, their usability for dispersion perturbations and their computational complexity. Meanwhile, the bit-error-rate (BER) floor in carrier phase estimation using a one-tap normalized LMS filter is evaluated analytically, and the numerical results are compared to a differential QPSK detection system. / QC 20110629

Every year wildfires threaten or destroy ecological habitats, man-made infrastructure and people’s lives. Additionally millions of dollars are spent each year trying to prevent and control these fires. Ideally if a wildfire can be detected before it rages out of control it can be extinguished and avoid large scale devastation. Traditional manned fire lookout towers are neither cost effective nor particularly efficient at detecting wildfire. It is proposed that temporal filtering can be used to isolate the signals created at the beginnings of potential wildfires. Temporal filtering can remove any background image and any periodic signals created by the camera movement. Once typical signals are analyzed, digital filters can be designed to pass fire signals while blocking the unwanted signals. The temporal filter passes only fire signals and signals generated by moving objects. These objects can be distinguished from each other by analyzing the objects mid and long wave energy profile. This algorithm is tested on 17 data sources and its results analyzed.

This thesis explores the challenge of automatic music transcription with a combination of digital signal processing and machine learning methods. Automatic music transcription is important for musicians who can't do it themselves or find it tedious. We start with an existing model, designed by Sigtia, Benetos and Dixon, and develop it in a number of original ways. We find that by using convolutional neural networks with filter shapes more tailored for spectrogram data, we see better and faster transcription results when evaluating the new model on a dataset of classical piano music. We also find that employing better practices shows improved results. Finally, we open-source our test bed for pre-processing, training, and testing the models to assist in future research.

In Laser Materials Processing there has always been a need for suitable methods to supervise and monitor the processes on line, to ensure correct production quality or to trigger alarms when failures are detected. Numerous investigations have been made in this field, including experimental and theoretical work. It is common practice in this field to monitor surface temperature, plasma radiation and back-reflected laser light, coaxially with the laser beam. Traditionally, the monitoring systems involved carry out no statistical analysis of the signals received - they merely involve thresholds. This thesis looks at the feedback collected during laser welding using such a co-axial setup from a Digital Signal Processing point of view and also uses high speed video photography to correlate signal perturbations with process anomalies.Modern Digital Signal Processing techniques such as Kalman filtering, Principal Component Analysis and Cluster Analysis have been applied to the measurement data and have generated new ways to describe the weld behaviour using parameters such as reflected pulse shape. The limitations of commercially available welding supervision systems have been studied and design suggestions for the next generation of on line weld monitoring equipment have been formulated.

No description available.

No description available.

This paper outlines the design of a high-precision graphic audio equalizer with digital filters in parallel, along with its implementation in MATLAB App Designer. The equalizer is comprised of 31 bands separated with a one-third octave frequency ratio, and its frequency response is controlled by 63 filters. Furthermore, the application can process audio signals, in real time, recorded by microphone and from audio files. While processing, it displays an FFT plot of the output sound, also in real time, equipped with a knob by which the refreshing pace can be adjusted. The actual frequency response proved to match the desired one accurately, but the matching is computationally demanding for the computer. An even higher accuracy would entail a computational complexity beyond the power of ordinary computers, and was thus concluded to be inappropriate. As a result, the final application manages to provide most laptops with both high precision and proper functionality.