Approved for public release, distribution is unlimited / Optical fiber has been the preferred cabling technology for certain building and campus network LAN backbones. Until recently, however, the use of fiber as a cabling medium to the desktop has been confined to special environments that require the unique properties of optical fiber such as noise immunity, security, distance, high bandwidth demands (CAD/CAM, video conferencing), and immunity to electrical interference. However, choosing to use optical fiber in a network over other cabling options may present significant advantages in its inherent ability to handle data at higher speeds. Decreasing costs of optical fiber components compared to the increasing electronic costs of carrying Gigabit Ethernet over Cat 5 or Cat 5E UTP copper cabling has also accelerated the migration to optical fiber LAN. The thesis conducts a feasibility study of implementing a Fiber Optic Local Area Network in Software Metrics Laboratory in Ingersoll 158. / Major, Republic of Singapore Air Force

Current coherent optical transmission systems focus on single carrier solutions for 400Gb/s serial transmission to support traffic growth in fiber-optics communications, together with a few subcarriers multiplexed solutions for the 1 Tb/s serial data rates and beyond. With the advancement of analog-to-digital converter technologies, high order modulation formats up to 64-QAM with symbol rate up to 72Gbaud have been demonstrated experimentally with Raman amplification. To enable such high serial data rates, it is highly desirable to implement in hardware low complexity digital signal processing schemes and advanced forward error correction coding with powerful error correction capability. In this dissertation, to enable high-speed optical communications, we first proposed an efficient FPGA architecture of high-performance binary and non-binary LDPC engines that can support throughputs of multiple Gb/s, which have low power consumption, providing high net coding gains at a target bit-error rate of 10-15. Further, we implement a generalized LDPC coding based rate adaptive binary LDPC coding scheme and puncturing based rate adaptive non-binary LDPC coding scheme, where large number of parameters can be reconfigured in order to cope with the time-varying optical channel conditions and service requirements. Based on comprehensive analysis on complexity, latency, and power consumption we demonstrate that the proposed efficient implementation represents a feasible solution for the next generation optical communication networks. Additionally, we investigate the FPGA implementation of rate adaptive regular LDPC coding combined with up to six high-order modulation formats and demonstrate high net coding gain performance and demonstrated a bit loading algorithm for irregular LDPC coding. Lastly, we present the real-time implementation of a direct detection OFDM transceiver with multi Giga symbols/s symbol rates in a back-to-back configuration.

Os sistemas de comunicação óptica por espaço livre, comumente denominados sistemas FSO (Free Space Optics), servem para estabelecer enlaces de comunicação do tipo wireless, ponto-a-ponto, a uma elevada taxa de dados e com alcances que podem variar de algumas centenas de metros a alguns quilômetros. O interesse pela tecnologia FSO tem sido estimulado pela necessidade de se fornecer soluções complementares às tradicionais (e.g. fibra óptica e RF) visando a atender à crescente demanda por conexões de banda larga. Algumas de suas aplicações mais importantes são a conexão de redes locais de computadores, a conexão de estações rádio base de telefonia móvel às centrais e o acesso de última milha. Embora a tecnologia FSO venha sendo empregada em pequena escala no Brasil, vislumbra-se o potencial de maior aplicação desta tecnologia em um futuro próximo, razão pela qual se propõe este estudo, focado no projeto de sistemas FSO. A principal contribuição deste trabalho é demonstrar uma possível solução para a realização de um enlace óptico operando em espaço livre, destinado a conectar duas redes locais de computadores Ethernet Rápida (100 Mbps). As seguintes atividades foram desenvolvidas: estudo das questões mais relevantes pertinentes ao projeto de sistemas FSO; projeto, construção e caracterização eletro-óptica de circuitos; montagem e demonstração do funcionamento de uma versão de testes do sistema, realizadas em laboratório; estudo de características associadas à transferência de radiação óptica de transmissor para receptor, envolvendo o balanço de potência para estimativa do alcance do enlace, sob condições atmosféricas ideais. Como resultado, obtivemos um sistema de comunicação do tipo full-duplex, operando no comprimento de onda de 850 nm em formato de modulação OOK, a uma taxa de dados efetiva de 125 Mbps. Com base na experiência obtida, melhoras ao sistema são sugeridas no final do trabalho. / Free space optical communication systems (FSO systems) provide wireless, point-to-point communication links at high data rates, at maximum distances ranging from hundreds of meters to several kilometers. The growing interest in FSO technology arises from the necessity of providing complementary and alternative solutions to those already adopted (eg. fiber optics and RF), aiming to satisfy the increasing demand for broadband connections. Some of its most important applications are connecting local area networks (LANs), backhauling wireless networks and providing last mile access. Although today FSO is not widely adopted in Brazil, the technology has the potential for being adopted in a greater scale in the near future, which is the reason we propose a study focused on issues related to the design of FSO systems. This work\'s main contribution is to demonstrate a possible solution for the construction of a FSO system intended to connect two Fast Ethernet LANs (100 Mbps). For this purpose, the following activities were developed: study of the main aspects related to FSO system design; design, construction and electrooptical characterization of circuits; mounting and demonstration, in the lab, of a test version of the system; study of characteristics related to radiation transfer from transmitter to receiver, involving the use of the link power budget for an estimation of the maximum distance achievable under ideal atmospheric conditions. The result is a full-duplex communication system that employs the 850 nm wavelength in OOK modulation format, operating at an effective bit rate of 125 Mbps. Based on the knowledge achieved, improvements to the system are suggested at the end of this work.

FundaÃÃo Cearense de Apoio ao Desenvolvimento Cientifico e TecnolÃgico / Serà estudado nesta dissertaÃÃo o comportamento da propagaÃÃo de trÃs tipos de pulsos Ãpticos (Solitons, Quasisolitons e Gaussianos) ultracurtos na regiÃo de 1.55μm e a eficiÃncia de conversÃo destes na geraÃÃo de pulsos por meio do fenÃmeno nÃo-linear de mistura de quatro ondas (FWM) em uma fibra comum e em fibras com dispersÃo decrescente em sistemas multiplexados por divisÃo de comprimento de onda (WDM). Para tal usamos seis diferentes perfis de dispersÃo; Constante, Linear, LogarÃtmico, Gaussiano, Exponencial e HiperbÃlico; nas simulaÃÃes da propagaÃÃo dos pulsos atravÃs de 100 Km de fibras.Na anÃlise dos resultados observamos que a eficiÃncia de conversÃo decai quando aumentamos a separaÃÃo entre os canais independentemente do perfil. Dentre estes perfis para os trÃs tipos de pulsos quando na separaÃÃo de canal de 0.2nm o HiperbÃlico apresentou a maior eficiÃncia, sendo de aproximadamente 1000 vezes maior em magnitude se comparada aos perfis Constante, Linear e LogarÃtmico.Para o pulso Gaussiano, exceto pelo perfil HiperbÃlico todos os outros perfis apresentaram a mesma eficiÃncia a partir da separaÃÃo de 0.675nm. Os pulsos Soliton e Quasisoliton tambÃm apresentam tal exceÃÃo no perfil Exponencial. A eficiÃncia de conversÃo ao longo do comprimento da fibra foi analisada para os trÃs tipos de pulsos em cada um dos seis perfis. Em todos os tipos de pulsos os perfis que apresentaram maior eficiÃncia de conversÃo, em ordem decrescente foram o HiperbÃlico, o Exponencial e o Gaussiano. Os outros trÃs, Constante, Linear e LogarÃtmico, apresentaram eficiÃncias muito prÃximas nos trÃs tipos de pulsos. Durante a propagaÃÃo observamos a presenÃa de pulsos gerados pela instabilidade modulacional para os trÃs tipos de pulsos nos perfis Constante, Linear, LogarÃtmico, Gaussiano e Exponencial. Entretanto apÃs percorrido aproximadamente 20Km a instabilidade modulacional desaparece nos trÃs primeiros perfis e nos outros dois o fenÃmeno permanece atà o final da propagaÃÃo na maioria dos casos. Em todos os casos estudados houve um aumento da eficiÃncia de conversÃo do pulso no canal 3 quando do incremento da potÃncia dos pulsos incidentes nos canais 1 e 4. A anÃlise do fator de compressÃo mostrou que os pulsos Soliton e Quasisoliton no canal 1 apresentaram alargamento na maioria dos perfis. E o pulso Gaussiano no canal 1 sà apresentou alargamento no perfil HiperbÃlico. No canal 4 o pulso Soliton sà apresentou compressÃo no perfil HiperbÃlico, o pulso Quasisoliton apresentou alargamento somente nos perfis Exponencial e Constante, no pulso Gaussiano todos os perfis apresentaram compressÃo, exceto o Constante que ao fim da propagaÃÃo obteve o mesmo valor que no inÃcio desta. A compressÃo do pulso gerado no canal 3 em todos os casos apresentou uma forte oscilaÃÃo nos primeiros 20 Km propagados e nos pulsos Soliton e Quasisoliton o perfil Gaussiano apresentou uma variaÃÃo brusca depois de percorrido 60 Km e estabilizando-se antes do fim da propagaÃÃo. / This dissertation will study the behavior of the propagation of three types of optical pulses (Solitons, Quasisolitons and Gaussianos, short ultra in the area of 1.55μm) and their efficiencies of conversion in the generation of pulses through the non linear phenomenon of four waves mixture (FWM) in a common fiber and in fibers with decreasing dispersion in systems multiplexados by wavelength division (WDM). Six different dispersion profiles were used during the simulations of propagation of the pulses through 100km of fibers: Constant, Linear, Logarithmic, Gaussiano, Exponential and Hyperbolic. The analysis of the results concluded that the efficiency of conversion declines when he/she increases the separation of the channels independently of the profile. The Hyperbolic profiles presented the largest efficiency when in the separation of channel of 0.2nm, being of approximately 1000 times larger in magnitude if compared to the other profiles Constant, Lineal and Logarithmic. In the pulse Gaussiano all profiles presented the same efficiency starting from the separation of 0.675nm, except for the Hyperbolic profile. While in the Soliton and Quasisoliton pulses, the Exponential profiel also presents such exception. The conversion efficiency along the length of the fiber was analyzed for the three types of pulses for each one of the six profiles. For all of the types of pulses the profiles that presented larger efficiency in decreasing order were, the Hyperbolic, the Exponential and Gaussiano. The other three, Constant, Lineal and Logarithmic, presented very close efficiencies for the three types of pulses. The presence of pulses generated by the instability modulacional for the three types of pulses in the profiles Constant, Lineal, Logarithmic, Gaussiano and Exponential was observed during the propagation. After travelling approximately 20Km the instability modulacional disappears in the first three profiles in all types of pulses while in the other two profiles the phenomenon stays until the end of the propagation in most of the cases. There was an increase of the efficiency of the conversion of the pulse in the channel 3 when of the increment of the potency of the incident pulses in the channels 1 and 4 in all the three types of pulses in each one of the six profiles. The analysis of the compression factor indicated that the pulse generated in the channel 3 by mixture of four waves (FWM) and the incident pulse in the channel 4 always presented compression during the propagation. The incident pulse in the channel 1 also presented compression in most of the profiles and in the Hyperbolic profile it presented strong enlargement attributed to the loss of energy. In the pulse generated in the channel 3 in all of the cases a strong oscillation was observed in the first 20 Km while in the Soliton and Quasisoliton pulses the Gaussian profile presented an abrupt variation after traveling 60 Km and being stabilized before the end of the propagation. This analysis concludes that the control of the efficiency of conversion and of the compression factor in the generation of mixture of four waves is made possible by the use of fiber optic with decreasing dispersion.

La demande en capacité liée à la transmission de tout type d’information (voix, vidéos, données, etc.) ne cesse de croître. Afin de répondre à cette demande croissante, de nouvelles générations de systèmes de communication multiplexés en longueur d’onde transmettant des débits élevés d’information par canal (100 Gbit/s ou plus) doivent être conçues. En plus des fibres ayant de très faibles pertes, des amplificateurs à fibre dopée à l’Erbium et du multiplexage en longueur d’onde, des technologies sont mises en place, comme notamment le multiplexage en polarisation, la détection cohérente, les formats de modulation multi-niveaux et plus récemment le multiplexage spatial. Des interrogations résident sur l’impact du multiplexage en polarisation ainsi qu’un développement vers des formats de modulation plus évolués incluant modulation de phase et multiplexage en polarisation. Dans cette thèse, afin de contribuer à l’augmentation du produit capacité x distance dans les systèmes de transmission Nx100 Gbit/s par fibre optique également multiplexés en polarisation et utilisant la détection cohérente, nous avons étudié d’une part, la mise en forme spectrale des signaux à l’émission pour augmenter la densité spectrale d’information (ISD: Information Spectral density). Dans cette optique, nous avons étudié l’impact du filtrage étroit gaussien du second ordre et de la mise en forme spectrale en racine de cosinus surélevé (RRC: Root Raised Cosine) sur les signaux émis dans le cas de modulations en Polarization Division Multiplexed-Quaternary Phase Shift Keying (PDM-QPSK) et Polarization-Switched-Quaternary Phase Shift Keying (PS-QPSK). Ceci a été réalisé en simulation numérique en considérant un espacement spectral entre les différents canaux variable. Nous avons montré qu’en tenant compte à la fois du facteur de qualité maximal et de la densité spectrale d’information, l’application de la mise en forme RRC sur des signaux modulés en PS-QPSK, fournit de meilleures performances de transmission dans une configuration où toute la dispersion est compensée en fin de propagation, pour toutes les valeurs d’espacement spectral étudiées. D’autre part, nous nous sommes intéressés aux effets non-linéaires qui limitent la portée de ces systèmes en dégradant pendant la propagation, les symboles émis, par les interactions entre des symboles d’un même canal, entre canaux ou modes de polarisation. La compréhension et la réduction de l’impact des effets non-linéaires est indispensable lorsqu’on veut utiliser certaines technologies pour augmenter la densité spectrale d’information. L’utilisation du multiplexage en polarisation par exemple, se heurte aux dégradations causées par les effets non-linéaires car de nouvelles interactions entre symboles sont présentes pendant la propagation. Par conséquent le développement des futurs systèmes ayant des débits plus élevés de 400 Gbit/s et 1 Tbit/s par canal passe par une diminution de l’impact des effets non-linéaires. Nous avons établi dans ce travail de thèse, des règles de conception permettant de réduire l’impact des effets non-linéaires entre polarisation dans les systèmes de transmission optiques considérés / The ever-increasing demand of capacity in very high bit rate coherent optical transmission systems has paved the way towards the investigation of several techniques such as the use of ultra-low loss fibers, Erbium doped fiber amplifiers, polarization and wavelength division multiplexing (WDM), coherent detection, multi-level modulation formats, spatial division multiplexing, etc. However, there are questions concerning polarization division multiplexing and a development towards some advanced modulation formats including phase modulation and polarization division multiplexing. In this thesis, in order to increase the capacity-by-distance product of future optical coherent systems using wavelength and polarization division multiplexing, we first study spectral shaping of the transmitted signals to increase the information spectral density. For this purpose, we have numerically investigated the multi-channel transmission performance of Polarization Switched Quadrature Phase Shift Keying (PSQPSK) and we have compared it to the performance of Polarization-Division-Multiplexed QPSK (PDM-QPSK), using Root Raised Cosine (RRC) spectral shaping, in the context of a flexible channel grid. In addition we have presented the advantage of PS-QPSK against PDM-QPSK as a function of the system parameters, while we have also discussed the benefit of a RRC spectral shaping against a tight filtering at the transmitter side with a 2nd order super-Gaussian-shaped filter. Furthermore, we have focused on nonlinear effects that limit the transmission distance by degrading the transmitted symbols during propagation. Analyzing and reducing the impact of nonlinear effects is essential when using technologies that increase the information spectral density such as polarization division multiplexing which causes new nonlinear effects due to additional interactions between symbols during the propagation through the fiber. Therefore a reduction of the impact of nonlinear effects is necessary for the development of future systems with higher bit rates of 400 Gbit/s and 1 Tbit/s per channel. We have established in this thesis, design rules to reduce the impact of nonlinear effects in the optical WDM transmission systems at 100 Gbit/s per channel that use polarization multiplexing

No description available.

Thesis (M.ICT.)--University of Wollongong, 2006. / Typescript. Includes bibliographical references: leaf 112-121.

Thesis (Ph. D.)--School of Computing and Engineering. University of Missouri--Kansas City, 2004. / "A dissertation in engineering and computer networking." Advisor: Ghulam Chaudhry. Typescript. Vita. Title from "catalog record" of the print edition Description based on contents viewed Feb. 22, 2006. Includes bibliographical references (leaves 135-157). Online version of the print edition.

Η παρούσα διατριβή εστιάζεται σε οπτικά τηλεπικοινωνιακά συστήματα με σύμφωνο δέκτη, τετραδική διαμόρφωση φάσης και πολύπλεξη κατά πόλωση. Το κυριότερο πλεονέκτημα αυτών των συστημάτων σε σύγκριση με τα συμβατικά οπτικά τηλεπικοινωνιακά συστήματα με δέκτες άμεσης φώρασης είναι οι αυξημένες δυνατότητες ψηφιακής επεξεργασίας σήματος στο δέκτη, που καθιστά εφικτή την εξίσωση των φαινομένων διάδοσης και την αποπολύπλεξη πόλωσης με ηλεκτρονικά μέσα. Μελετώνται διατάξεις ηλεκτρονικής επεξεργασίας σήματος. Η μελέτη διεξάγεται με χρήση προσομοίωσης με τη βοήθεια ηλεκτρονικού υπολογιστή. Προτείνεται η χρήση δύο πρωτότυπων μεθόδων αποπολύπλεξης σημάτων πολυπλεγμένων κατά πόλωση, με πλεονεκτήματα ως προς την απόδοση, την αξιοπιστία και την ταχύτητα σύγκλισης σε σύγκριση με τις μεθόδους που έχουν προταθεί στη βιβλιογραφία Στη συνέχεια μελετάται η απόδοση των ηλεκτρονικών εξισωτών ελαχίστου μέσου τετραγώνου (least-mean square) και σταθερής περιβάλλουσας (constant modulus algorithm) στην εξίσωση της χρωματικής διασποράς και της διασποράς τρόπων πόλωσης, τόσο μεμονωμένα, όσο και από κοινού. Στη μελέτη μας συμπεριλαμβάνουμε το θόρυβο φάσης και την ενδιάμεση συχνότητα, αίροντας αντίστοιχες απλοποιήσεις που συνήθως γίνονται στη βιβλιογραφία. Ακολούθως αναφερόμαστε στο πρόβλημα του εκφυλισμού του πίνακα συντελεστών των εξισωτών ελαχίστου μέσου τετραγώνου και σταθερής περιβάλλουσας παρουσία έντονης διεξασθένησης. O πίνακας συντελεστών στον οποίο έχει συγκλίνει ο εξισωτής ονομάζεται εκφυλισμένος όταν παράγεται η ίδια απεσταλμένη ακολουθία συμβόλων και στις δύο εξόδους του εξισωτή. Προτείνουμε τη χρήση του πολυχρηστικού εξισωτή σταθερής περιβάλλουσας (multiuser constant modulus algorithm equalizer). Η λύση που προτείνεται έχει πλεονεκτήματα αξιοπιστίας και απλότητας σε σχέση με αντίστοιχες προτάσεις στη βιβλιογραφία, και αποδεικνύεται η άριστη απόδοσή της. Στη συνέχεια, γίνεται αναφορά στο φαινόμενο της ανισοσταθμίας ορθογωνιότητας που παρατηρείται στα οπτικά συστήματα με σύμφωνους δέκτες και αποτελεί ένα φαινόμενο το οποίο, κατά κανόνα, παραβλέπεται στις μελέτες της βιβλιογραφίας και αποδεικνύεται το ότι οι συμβατικοί ηλεκτρονικοί εξισωτές δεν μπορούν να το αντιμετωπίσουν. Τέλος, παραθέτουμε αποτελέσματα από την επεξεργασία πραγματικών πειραματικών δεδομένων, για την επαλήθευση των συμπερασμάτων που προέκυψαν θεωρητικά. / This PhD dissertation focuses on optical communication systems utilizing coherent receiver, quadrature phase modulation and polarization multiplexing. Their main advantage compared to conventional systems is the ability to perform extended digital signal processing, in order to compensate for transmission effects and to successfully demultiplex polarization multiplexed signals. We study such DSP modules with the help of computer simulations. We propose two novel polarization demultiplexing algorithms, with advantages in performance reliability and convergence speed compared to conventional methods. One of the two methods utilizes training procedure, an approach not yet studied enough in the literature. Next, we study the performance of least mean square and constant modulus algorithm equalizers in mitigating chromatic dispersion and polarization mode dispersion, both separately and in combination. We include in our study phase noise and intermediate frequency offset, without the simplifications usually found in the literature. Next, we address the subject of equalizer degeneration, which describes the phenomenon when both equalizer outputs transmit one of the two polarization multiplexed signals. We propose the use of multi-user constant modulus algorithm, which is an enhanced version of the conventional constant modulus algorithm. We prove that our solution has excellent performance as well as reliability and simplicity advantages compared to other methods. Next, we prove that quadrature imbalance is a ubiquitous phenomenon that cannot be addressed with conventional electronic equalizers and should not be omitted as is often the case. Finally, we present results derived from the application of some of the proposed algorithms on real experimental data, and prove the consistency of the theoretical derived results.

Orientadores: Evandro Conforti, Carlos Allan Caballero Petersen / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-16T07:18:34Z (GMT). No. of bitstreams: 1 Toazza_AdrianoLuis_D.pdf: 4079607 bytes, checksum: bb3aa0dd52628e849f97cb827da20ffb (MD5) Previous issue date: 2010 / Resumo: A técnica de pré-distorção com a adição de um impulso ao degrau que normalmente chaveia o Amplificador Óptico a Semicondutor - (SOA) , permite diminuir os tempos de chaveamento eletro-óptico do SOA para valores inferiores a nanosegundos, obtendo-se contrastes da luz amplificada na saída da chave acima de 25 dB e bandas de passagem da ordem de 60 nm. Entretanto, estes tempos de chaveamento de alta velocidade vêm acompanhados de efeitos deletérios ligados às flutuações do ganho óptico durante o processo de ligar a chave, do estado "off" para o estado "on". Nesta tese é efetuada a análise destes fenômenos, com boa correspondência com resultados experimentais. Mostram-se, também, caminhos a seguir para diminuir as flutuações de ganho. O estudo aqui apresentado baseia-se no comportamento da impedância da montagem com o SOA em alta frequência, até 20 GHz. A obtenção dos valores dos componentes do circuito equivalente incluem o SOA (a partir de sua porta elétrica) e a montagem de microondas do chaveamento eletro-óptico. O modelo para o SOA foi baseado em um circuito desenvolvido para laser a diodo semicondutor. Os parâmetros do circuito equivalente para a montagem do SOA foram obtidos por técnicas de extração de parâmetros, através de aproximações sucessivas entre as respostas experimentais e teóricas, utilizando um programa desenvolvido neste trabalho. Em seguida, a partir da impedância do SOA, foram estimados os tempos de transição, o casamento em banda larga, assim como a influência do encapsulamento do dispositivo, aprimorando o entendimento do seu comportamento e limitações, com boa correspondência com resultados experimentais / Abstract: Reduction of the Semiconductor Optical Ampli.er (SOA) switching times can be achieved with the pre-distortion technique consisting of impulse(s) addition to the current step that generally switches the SOA. With this technique it is possible to reduce the electro-optical switching time to sub-nanoseconds with a contrast approaching 25 dB between the "off" and the "on" state of the optical gain, and with a bandwidth in excess of 60 nm. However those high speed switching times comes jointly with deleterious effects of overshoot and optical gain fluctuations during the switch state variation from the "off" to the "on" state. This thesis analyzes those phenomena with good correspondence to experimental results, and actions to decrease those deleterious effects are revealed. The results make use of SOA electrical gate impedance measurements up to 20 GHz, with the accomplishment of the SOA equivalent circuit extraction including the effects of the microwave lines and SOA encapsulation. The SOA model was based in the literature results for semiconductor lasers. The parameters extraction was based in a software and in measurements prepared here to obtain the circuit parameters values through successive approximations between practice and theory. Finally, the obtained SOA equivalent circuit was used to find the switching time estimation, the broadband matching, the encapsulation influences, in order to achieve a better theoretical understanding of the deleterious effects. The computer simulated results are in good correspondence with the experimental results / Doutorado / Telecomunicações e Telemática / Doutor em Engenharia Elétrica