Σε αυτή την εργασία μελετάμε τους αλγορίθμους που χρησιμοποιούνται στην αποκωδικοποίηση των LDPC, με έμφαση στους κώδικες του προτύπου 802.11n. Αντιμετωπίζουμε τις δυσκολίες που αντιμετωπίζουν στην υλοποίηση στο υλικό, κυρίως στην εκτέλεση αριθμητικών πράξεων, και προτείνουμε πρακτικές λύσεις. Χρησιμοποιώντας τα αποτελέσματα εκτενών εξομοιώσεων, καταλήγουμε στις βέλτιστες παραμέτρους που θα έχουν οι προτεινόμενες υλοποιήσεις. / In this thesis, we study the LDPC decoding algorithms, with emphasis on the 802.11n standard codes. We tackle the hardware implementation difficulties, especially those related to arithmetic computations, and propose practical solutions. Leveraging the results of extensive simulations, we find the optimal parameters of the proposed implementations.

Σχεδόν όλα τα σύγχρονα τηλεπικοινωνιακά συστήματα, τα οποία προορίζονται για αποστολή δεδομένων σε υψηλούς ρυθμούς, έχουν υιοθετήσει κώδικες διόρθωσης λαθών για την αύξηση της αξιοπιστίας και τη μείωση της απαιτούμενης ισχύος εκπομπής τους. Μια κατηγορία κωδίκων, και μάλιστα με εξαιρετικές επιδόσεις, είναι η οικογένεια των LPDC κωδίκων (Low-Density-Parity-Check codes). Οι κώδικες αυτοί είναι γραμμικοί block κώδικες με απόδοση πολύ κοντά στο όριο του Shannon. Επιπλέον, ο εύκολος παραλληλισμός της διαδικασίας αποκωδικοποίησής τους, τους καθιστά κατάλληλους για υλοποίηση σε υλικό. Στην παρούσα διπλωματική μελετούμε τα ιδιαίτερα χαρακτηριστικά και τις παραμέτρους των κωδίκων αυτών, ώστε να κατανοήσουμε την εκπληκτική διορθωτική ικανότητά τους. Στη συνέχεια, επιλέγουμε μια ειδική κατηγορία κωδίκων LDPC, της οποίας οι πίνακες ελέγχου ισοτιμίας έχουν δημιουργηθεί ώστε να διευκολύνουν την υλοποίησή τους, και προχωρούμε στο σχεδιασμό αυτής σε υλικό. Πιο συγκεκριμένα, υλοποιούμε σε VHDL έναν αποκωδικοποιητή σύμφωνα με τον rate ½ και block_lenght 576 bits πίνακα του προτύπου WiMax 802.16e, με στόχο κυρίως την επίτευξη πολύ υψηλού throughput. Στο χρονοπρογραμματισμό της μετάδοσης των μηνυμάτων μεταξύ των κόμβων του κυκλώματος χρησιμοποιούμε το two-phase scheduling και προτείνουμε δύο τροποποιήσεις αυτού για την επιτάχυνση της διαδικασίας αποκωδικοποίησης, οι οποίες καταλήγουν σε 24 και 50% βελτίωση του απαιτούμενου χρόνου μιας επανάληψης με μηδενική και σχετικά μικρή αύξηση της επιφάνειας ολοκλήρωσης αντίστοιχα. Ο όλος σχεδιασμός είναι πλήρως συνθέσιμος και η σωστή λειτουργία αυτού έχει επιβεβαιωθεί σε επίπεδο λογικής εξομοίωσης. Κατά τη διάρκεια σχεδιασμού, χρησιμοποιήθηκαν εργαλεία της Xilinx και MentorGraphics. / Αlmost all the modern telecommunication systems, which are designed for high data rate transmissions, have adopted error correction codes for improving the reliability and the required power of transmission. One special group of these codes, with extremely good performance, is the LDPC codes (Low-Density-Parity-Check codes). These codes are linear block codes with performance near to the theoretical Shannon limit. Furthermore, the inherent parallelism of the decoding procedure makes them suitable for implementation on hardware. In this thesis, we study the special characteristics of these codes in order to understand their astonishing correcting capability. Then, we choose a special category of these codes, whose parity check matrix are special designed to facilitate their implementation on hardware, and we design a high-throughput decoder. More specifically, we implement in VHDL an LDPC decoder according to the rate ½ and block_length 576 bits code of WiMax IEEE802.16e standard, with main purpose to achieve very high throughput. We use the two-phase scheduling at the message passing and we propose 2 modifications for reducing the required decoding time, which result in 25 and 50% improving of the required decoding time of one iteration with zero and little increasing in the decoder’s area respectively. Our design has been successfully simulated and synthesized. During the design process, we used Xiinx and MentorGraphics’s tools.

No description available.

Orientador: Celso de Almeida / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-12T12:24:43Z (GMT). No. of bitstreams: 1 OlmedoCifuentes_GonzaloFernando_D.pdf: 1566286 bytes, checksum: 91619aa7cb33f0594cf4dd4e9b185238 (MD5) Previous issue date: 2008 / Resumo: Neste trabalho foi proposto um novo algoritmo para o controle da janela de congestionamento do protocolo TCP, que melhora o desempenho da vazão de dados em redes sem fio. A estimação da banda disponível responsável pela melhoria do sistema, se dá pois os segmentos TCP com erro produzidos pelo canal sem fio são reconhecidos por uma confirmação negativa NACK, ou quando o buffer do sistema enche. Para o mecanismo proposto, uma expressão analítica da vazão de dados em função da probabilidade de erro de segmentos TCP no canal sem fio é obtida. No sistema sem fio é analisado o enlace reverso de sistemas CDMA de sequencia direta, tanto síncrono, como assíncrono, considerando um canal seletivo em frequencia com desvanecimento do tipo Rayleigh e modulação BPSK. São comparados diferentes tipos de receptores, que usam detecção multiusuário descorrelacionadora, arranjo de antenas e codificação de canal do tipo convolucional ou turbo. Foram obtidas expressões simples e precisas, antes não encontradas na literatura da eficincia assintotica média e da probabilidade de erro de bit que foram comparadas com resultados de simulação. Foram comparadas tres estratégias propostas de decodificação iterativa turbo. Expressões de complexidade computacional do receptor também foram elaboradas. / Abstract: In this work, we propose a new TCP congestioned window algorithm that improves the performance for wireless networks, where the TCP segments with errors caused by the wireless channel are recognized by a negative acknowledgement (NACK) or when the system buffer is full, the protocol estimate the available bandwidth. For the proposed strategy, a simple analytic characterization of the steady state throughput as a function of the TCP segment error probability for the wireless channel is obtained. In the wireless system, we consider the reverse link of synchronous and asynchronous direct-sequences code division multiple access (DS-CDMA) systems in frequency-selective Rayleigh fading channels and BPSK modulation. Different types of receivers using multiuser decorrelating detection, antenna array and convolutional or turbo coding are compared. Simple and precise novel close-form analytic expressions for the asymptotic multiuser efficiency and for the mean bit error probability are derived. The proposed expressions of the analytic model developed to evaluate the performance of those systems are verified by the results of computer simulations. For turbo code, three iterative turbo decoding strategies are proposed and compared. Expressions for computational complexity are also obtained. / Doutorado / Telecomunicações e Telemática / Doutor em Engenharia Elétrica

No description available.

The main objective of this thesis was to create a model of forward error correction (FEC) system used in ADSL modem, which will be able to check its individual settings using the test methods specified in Recommendation G.996.1. The work is divided into two basic parts – theoretical and practical. The first part describes the basic functional characteristics of ADSL system, with attention being paid to aspects affecting the activity and setting of the error protection system. There are also discussed the theoretical basics of FEC coding techniques used in ADSL. The work also contains a description of the method for testing DSL system resistance against impulsive interference, which is specified in Recommendation G.996.1. The practical part of this thesis contains a description of the proposed model of the FEC system. The model was implemented in the programming environment Matlab Simulink. The proposed model includes the simulation of FEC coding process, DMT modulation and transmission channel with the impulse interference and the background noise. The proposed program is able to calculate parameters for setting FEC system from the specified data transfer rates and the requirement for interleaving. In addition, allows you to set the parameters of transmission line and sources of impulse interference and of the background noise. Under the set of input characteristics of the transmission is calculated bit allocation and the resulting parameters are given to a computational core of the model. In the application can be performed two types of test tasks. The first one is based on the Recommendation G.996.1 and makes it possible to determine probability of errored second. The second test task is for the measurement of resistance of the transmission system against impulsive interference. The results of the simulations are summarized in the conclusion of work.

Master's thesis compares the standard for digital terrestrial television broadcasting of the first generation DVB-T/H and the second generation DVB-T2 with particular emphasis on the modulator, a security channel interference, the signal from the transmitting environment itself and the modulation signal. The following description of specific models of transmission channels for fixed, portable and mobile reception of digital terrestrial signal. Comparison with the particular relates of the transmission parameters for different types of reception of digital terrestrial television with achieving efficiency and effectiveness of transmission at the level of laboratory measurements and computer simulation.

A oferta de novos serviços para os usuários finais, como o denominado triple play, que consiste no tráfego simultâneo de voz, vídeo e dados utilizando a mesma infra-estrutura de comunicação, vem exigindo que as estruturas de rede das operadoras ofereçam largura de banda adequada e qualidade de serviço. Nesse contexto, as redes ópticas passivas (PON) vêm se destacando em virtude de oferecerem maior largura de banda a custos relativamente baixos. Nas redes ópticas passivas, trechos de fibras ópticas podem ser compartilhados entre diversos assinantes, exigindo, para isso, a utilização de técnicas de controle de acesso múltiplo. Destaque maior é dado à técnica de acesso múltiplo por divisão de códigos ópticos (OCDMA), por apresentar características tais como maior segurança e capacidade flexível sob demanda. O desempenho dessa tecnologia é basicamente limitado pela interferência de acesso múltiplo, ou interferência multiusuário (MAI). No presente trabalho, cenários OCDMA-PON utilizando códigos ópticos unidimensionais, baseados na codificação prima modificada (MPC), e bidimensionais, baseados na codificação óptica ortogonal de múltiplos comprimentos de onda (MWOOC), são descritos, e seus respectivos desempenhos investigados. Os desempenhos desses sistemas são verificados utilizando-se os seguintes esquemas de modulação: 1) On-off Keying (OOK), cujo formalismo para a codificação padded MPC (PMPC) e double padded MPC (DPMPC) foi aqui desenvolvido, 2) por posição de pulso (PPM), considerando os ruídos poissonianos de deteção, cujo formalismo foi também aqui desenvolvido, 3) por posição de pulso (PPM) considerando apenas a interferência de acesso múltiplo (MAI), 4) semelhante ao (3), porém, utilizando a técnica mitigadora de MAI via cancelamento paralelo de interferência (PIC), e 5) semelhante ao (4) porém utilizando modulação por chaveamento de frequência (FSK). Em seguida, a codificação bidimensional MWOOC, utilizando modulação OOK, é analisada com e sem a utilização de PIC. Além desses esquemas, são também incluídas, pela primeira vez nesses formalismos, técnicas de correção posterior de erro (FEC) baseadas no algoritmo Reed-Solomon (RS). A figura de mérito adotada nas simulações é baseada no cálculo da taxa de erro de bit (BER). Os resultados obtidos para os diversos casos investigados neste trabalho indicam que a combinação PIC/FEC deve ser considerada de modo a se obter níveis de BER compatíveis com a região livre de erros (BER < \'10 POT.-12\') para os cenários de redes de acesso atuais e futuros (visando atender a 32 ou 64 usuários simultâneos). / The advent of new services offered to end users, such as the so called triple play, which consists in the simultaneous traffic of voice, video, and data through the same communication infrastructure, has been pushing telecom operators towards providing adequate bandwidth as well as quality of service. This has brought the spotlight to passive optical networks (PONs) by virtue of their considerably higher bandwidth at relatively low cost. In PONs, certain fiber optic spans can be shared among different users as long as multiple access control techniques are used. One such technique is the optical code division multiple access (OCDMA) technique, which, in addition, presents improved data security and flexible capacity on demand. The performance of this technology is basically limited by multi access interference (MAI), which becomes more severe as the number of simultaneous users increases. In the present work, OCDMA-PON scenarios utilizing one- and two-dimensional optical codes separately such as modified prime codes (MPC) and multi-wavelength optical orthogonal codes (MWOOC), are described and thoroughly investigated. The performances of these systems are investigated based on the following modulation schemes: 1) On-off keying (OOK), where a new formalism for the double-padded MPC is proposed, 2) Pulse position modulation (PPM), and 3) Frequency shift keying (FSK) together with MAI mitigation technique based on parallel interference cancelation (PIC). Besides these schemes, it is also included for the first time in these formalisms forward error correction techniques (FEC) based on the Reed-Solomon (RS) algorithm. The figure of merit adopted in all simulations is the bit error rate (BER). The results obtained for the cases investigated in this work indicate that the combination PIC/FEC must be considered if one is interested in obtaining BER levels compatible with the error-free region (BER < \'10 POT.-12\') for the current and future access network scenarios (aiming at attending 32 or 64 simultaneous users).

A oferta de novos serviços para os usuários finais, como o denominado triple play, que consiste no tráfego simultâneo de voz, vídeo e dados utilizando a mesma infra-estrutura de comunicação, vem exigindo que as estruturas de rede das operadoras ofereçam largura de banda adequada e qualidade de serviço. Nesse contexto, as redes ópticas passivas (PON) vêm se destacando em virtude de oferecerem maior largura de banda a custos relativamente baixos. Nas redes ópticas passivas, trechos de fibras ópticas podem ser compartilhados entre diversos assinantes, exigindo, para isso, a utilização de técnicas de controle de acesso múltiplo. Destaque maior é dado à técnica de acesso múltiplo por divisão de códigos ópticos (OCDMA), por apresentar características tais como maior segurança e capacidade flexível sob demanda. O desempenho dessa tecnologia é basicamente limitado pela interferência de acesso múltiplo, ou interferência multiusuário (MAI). No presente trabalho, cenários OCDMA-PON utilizando códigos ópticos unidimensionais, baseados na codificação prima modificada (MPC), e bidimensionais, baseados na codificação óptica ortogonal de múltiplos comprimentos de onda (MWOOC), são descritos, e seus respectivos desempenhos investigados. Os desempenhos desses sistemas são verificados utilizando-se os seguintes esquemas de modulação: 1) On-off Keying (OOK), cujo formalismo para a codificação padded MPC (PMPC) e double padded MPC (DPMPC) foi aqui desenvolvido, 2) por posição de pulso (PPM), considerando os ruídos poissonianos de deteção, cujo formalismo foi também aqui desenvolvido, 3) por posição de pulso (PPM) considerando apenas a interferência de acesso múltiplo (MAI), 4) semelhante ao (3), porém, utilizando a técnica mitigadora de MAI via cancelamento paralelo de interferência (PIC), e 5) semelhante ao (4) porém utilizando modulação por chaveamento de frequência (FSK). Em seguida, a codificação bidimensional MWOOC, utilizando modulação OOK, é analisada com e sem a utilização de PIC. Além desses esquemas, são também incluídas, pela primeira vez nesses formalismos, técnicas de correção posterior de erro (FEC) baseadas no algoritmo Reed-Solomon (RS). A figura de mérito adotada nas simulações é baseada no cálculo da taxa de erro de bit (BER). Os resultados obtidos para os diversos casos investigados neste trabalho indicam que a combinação PIC/FEC deve ser considerada de modo a se obter níveis de BER compatíveis com a região livre de erros (BER < \'10 POT.-12\') para os cenários de redes de acesso atuais e futuros (visando atender a 32 ou 64 usuários simultâneos). / The advent of new services offered to end users, such as the so called triple play, which consists in the simultaneous traffic of voice, video, and data through the same communication infrastructure, has been pushing telecom operators towards providing adequate bandwidth as well as quality of service. This has brought the spotlight to passive optical networks (PONs) by virtue of their considerably higher bandwidth at relatively low cost. In PONs, certain fiber optic spans can be shared among different users as long as multiple access control techniques are used. One such technique is the optical code division multiple access (OCDMA) technique, which, in addition, presents improved data security and flexible capacity on demand. The performance of this technology is basically limited by multi access interference (MAI), which becomes more severe as the number of simultaneous users increases. In the present work, OCDMA-PON scenarios utilizing one- and two-dimensional optical codes separately such as modified prime codes (MPC) and multi-wavelength optical orthogonal codes (MWOOC), are described and thoroughly investigated. The performances of these systems are investigated based on the following modulation schemes: 1) On-off keying (OOK), where a new formalism for the double-padded MPC is proposed, 2) Pulse position modulation (PPM), and 3) Frequency shift keying (FSK) together with MAI mitigation technique based on parallel interference cancelation (PIC). Besides these schemes, it is also included for the first time in these formalisms forward error correction techniques (FEC) based on the Reed-Solomon (RS) algorithm. The figure of merit adopted in all simulations is the bit error rate (BER). The results obtained for the cases investigated in this work indicate that the combination PIC/FEC must be considered if one is interested in obtaining BER levels compatible with the error-free region (BER < \'10 POT.-12\') for the current and future access network scenarios (aiming at attending 32 or 64 simultaneous users).

ITC/USA 2011 Conference Proceedings / The Forty-Seventh Annual International Telemetering Conference and Technical Exhibition / October 24-27, 2011 / Bally's Las Vegas, Las Vegas, Nevada / The Central Test and Evaluation Investment Program (CTEIP) Integrated Network Enhanced Telemetry (iNET) program is currently testing networked telemetry transceivers (IP.TM-Tx/Rx) using the Internet Protocol (IP), for use in telemetry (TM) channels. A unique characteristic of networked telemetry channel is packet drops due to radio frequency (RF) signal dynamics, i.e., terrain, weather, aircraft attitude, manmade objects, etc.. One of the key measures of the IP.TMTx/ Rx is reliability is link availability (LA), and a key element of LA is time to resynchronize after RF link loss.