<p>To let many users benefit from the high speed communication in an optical fiber, a way to share the channel is needed. One way to share the channel between the users is the multiple access technique of Code Division Multiple Access, CDMA. This technique is based on coding. By assigning the users different codewords, the receiver is able to distinguish the information sent from each user. One family of codes suitable to use in a fiber optic network is the constant weight Optical Orthogonal Codes, OOC.</p><p>To influence the users interference tolerability it is possible to assign the users codewords of the same length but with different weights. By letting the users have codewords of different length it is possible to have different data rates in the system. These two versions of optical orthogonal codes, with variable length and weight, are investigated in this thesis for their cardinality. The variable weight and variable length codes were created by combining optical orthogonal codes with different parameters.</p><p>I have found that for variable length OOC:s of weight 2 it seems to be possible to increase the cardinality of a code, but for codes with higer weights than that, it is better that all users are assigned codewords of the same length. But since an upper bound for the cardinality of these types of codes has not been found, it is not possible to verify if this conclusion is correct. For variable weight OOC:s it is found that it is only possible to increase the cardinality of small, not optimal, codes. For codes including many codewords it is rarely possible to include more codewords without breaking the conditions on cross correlation between the codewords.</p>

To let many users benefit from the high speed communication in an optical fiber, a way to share the channel is needed. One way to share the channel between the users is the multiple access technique of Code Division Multiple Access, CDMA. This technique is based on coding. By assigning the users different codewords, the receiver is able to distinguish the information sent from each user. One family of codes suitable to use in a fiber optic network is the constant weight Optical Orthogonal Codes, OOC. To influence the users interference tolerability it is possible to assign the users codewords of the same length but with different weights. By letting the users have codewords of different length it is possible to have different data rates in the system. These two versions of optical orthogonal codes, with variable length and weight, are investigated in this thesis for their cardinality. The variable weight and variable length codes were created by combining optical orthogonal codes with different parameters. I have found that for variable length OOC:s of weight 2 it seems to be possible to increase the cardinality of a code, but for codes with higer weights than that, it is better that all users are assigned codewords of the same length. But since an upper bound for the cardinality of these types of codes has not been found, it is not possible to verify if this conclusion is correct. For variable weight OOC:s it is found that it is only possible to increase the cardinality of small, not optimal, codes. For codes including many codewords it is rarely possible to include more codewords without breaking the conditions on cross correlation between the codewords.

De nos jours, les liaisons par fibres optiques monomode ou multimodes équipent de plus en plus de bâtiments, de lieux et sites de transports publics (gares, aéroports, autoroutes, ports ou plateformes multimodales). Ceci est dû au débit important qu’offre la fibre optique et à son atténuation très faible sur de longues distances. Afin de gérer au mieux la transmission au sein de la fibre optique, les performances de divers codes à une dimension 1D OOC et PC et deux dimensions 2D-MWOOC sur les récepteurs conventionnels ont été étudiées. Cependant, les performances obtenues restent limitées lorsque le nombre d’utilisateurs croit significativement. En ne tenant pas compte du bruit causé par les composants du système, la dégradation des performances est principalement due aux interférences d’accès multiples (IAM). De ce fait, nous proposons dans le cadre de cette thèse d’optimiser les performances du récepteur PIC en ajoutant un facteur de compensation qui permet de réduire au mieux les interférences d’accès multiples. Ceci est étudié avec les codes optiques à deux dimensions 2D-PC/PC que nous générons. Des expérimentations sur un banc d’essai optique en laboratoire sont effectuées pour illustrer la mise en oeuvre de multiservices. Une validation de certains résultats théoriques est également menée sur un simulateur dédié. / Nowadays, public transport spaces (train stations, airports, highways, ports or multimodal platforms) and buildings are being more and more equipped by single mode or multimode optical fibers. The high throughput offered and the very low attenuation over long distances proposed by optical fiber are indeed very promising. For a better management of optical fiber transmissions, the performance of various one-dimensional 1D OOC and PC codes, and twodimensional2D- MWOOC codes on conventional receivers have been previously studied.These studies show that a good level of performance is only available to a restricted number of users. Assuming noiseless components in the system, the performance degradation is mainly due to multiple access interference (MAI). Therefore, we propose in this thesis to optimize the performance of PIC receivers by adding a compensation factor that can best reduce the MAI. This is studied with two-dimensional optical codes 2D-PC/PC that we have generated. Experiments on a laboratory test bench were made to implement multiple services. The theoretical studies and the channel optical transmission simulations were also validated by using a dedicated simulator tool.

[Abstract]: The disposal of animal waste has become a problem in many parts of the world due to the rapid growth in the number and the size of intensive animal industries. Safe waste disposal sites are rarely available and the relocation and/or treatment of animal waste is seldom economically viable. The reuse of animal waste for energy recovery and re-feeding is also not popular. Animal waste is a valuable source of plant nutrients and a very good soil conditioner, and has been commonly applied as fertiliser to agricultural fields. However, due to the increasing oversupply of animal waste in recent years, it has often been applied in excess to the agricultural fields. Excessive application of animal waste, without due consideration of its implications, is a serious concern. The run-off and leaching losses of nutrients from the fields fertilised with animal waste have contributed significantly to the eutrophication and toxic blue-green algae blooms in surface water systems and nitrification of ground water systems. It has also led to nutrient imbalances in the soils and odour pollution to the surrounding communities. The animal waste, which is a valuable source of plant nutrients, has thus become both an economic and environmental burden, and there is a need to develop a strategy for its sensible use as a fertiliser in agricultural fields. Sensible use of animal waste involves the consideration of all the agricultural, environmental, social, and economical limitations. A rational method of achieving this is to restrict the use of animal waste to sites suitable for such uses, identify areas where it can be relocated and applied economically, limit the application rates to a safe level, and observe appropriate manure management practices. This study addressed each of these components by developing a comprehensive manure application plan (MAP) for the site-specific use of animal waste as fertiliser in agricultural fields. Various geographic information systems (GIS) based techniques, including a weighted linear combination model and map algebra based cartographic modelling, were employed to achieve the goal. The appropriateness of the existing techniques and procedures were evaluated and modified to meet the current input requirements. New methods of analysis were devised as necessary. The Westbrook sub-catchment of the Condamine River catchment in south-east Queensland was selected as the study area. The sub-catchment covers 24,903 hectares and contains 39 intensive animal industries. The catchment is also a part of the Murray-Darling Basin, which has been suffering from toxic blue green algae blooms recurrently since 1991. This study identified that only about one-fifth of the sub-catchment area is suitable for animal waste application. Depending on the method of site suitability analysis and the number of input factors used the suitable area ranged between 16 and 22 percent. This comparatively small area is mainly due to the presence of a large proportion of non-agricultural areas in the sub-catchment. The suitable areas were also found to have various degrees of suitability for waste application. However, the degree of site suitability was affected by the number of input factors used in the analysis, the weighting of the factors, and the method of factor attribute standardisation. Conventional methods of weighting input factors were found to be cumbersome and not particularly suitable. Hence, this study developed a new ‘objective oriented comparison’ method of factor weighting. Standardisation of input factors using a continuous, rather than discrete, classification (ie fuzzy set) method was found to be more consistent in degree of suitability determination. The discrete classification of factor attributes into classes of different numbers and sizes, and the weighting of classes to a sum of one, were identified as a limitation in using this standardisation method. A new ‘weight adjustment’ method was devised and demonstrated to reduce factor-weighting biases. The suitable sites, degree of site suitability, and other relevant spatial and non-spatial information were processed within a GIS framework to develop a comprehensive manure application plan. The inherently high presence of available phosphorus in the soils of the study area was recognised and the P2O5 content in the manure was used as the basis for determining manure application rates. A complimentary nitrogen supply map was also generated. Manure management practices applicable to the areas with a lower degree of suitability were also suggested.

<p>To take advantage of the high speed in an optic fiber, one of the basic concept in fiber optic communication is to allow several users to simultaneously transmit data over the channel. One technique that provides multiple access is it fiber optic-code division multiple access (FO-CDMA). In FO-CDMA each user is assigned one or more signature sequences called codewords, which are subsets of a type of optical orthogonal code (OOC). The channel input/output consists of the superposition of several users codewords and at the receiver end an optical correlator extracts the information. </p><p>In the parallel code constructions, presented in this report, each user j is assigned a subset Cj from a code C. The subsets are disjoint and their union is the whole set C. A new way to map the information bits is to insert up to L zeros before each codeword from Cj and let this represent information aswell. This gives high rates for active users but an investigation is needed to ensure that this does not compromise the systems wanted property of sending information with a small probability of errors for all users. Therefore a simulation environment has been implemented in Matlab. </p><p>The result from these simulations shows that BER for the L parallel codes is acceptable and not much higher than for the traditional constructions. Because of the higher rate these construction should be preferred but an analysis if a hardware implementation is possible.</p>

To take advantage of the high speed in an optic fiber, one of the basic concept in fiber optic communication is to allow several users to simultaneously transmit data over the channel. One technique that provides multiple access is it fiber optic-code division multiple access (FO-CDMA). In FO-CDMA each user is assigned one or more signature sequences called codewords, which are subsets of a type of optical orthogonal code (OOC). The channel input/output consists of the superposition of several users codewords and at the receiver end an optical correlator extracts the information. In the parallel code constructions, presented in this report, each user j is assigned a subset Cj from a code C. The subsets are disjoint and their union is the whole set C. A new way to map the information bits is to insert up to L zeros before each codeword from Cj and let this represent information aswell. This gives high rates for active users but an investigation is needed to ensure that this does not compromise the systems wanted property of sending information with a small probability of errors for all users. Therefore a simulation environment has been implemented in Matlab. The result from these simulations shows that BER for the L parallel codes is acceptable and not much higher than for the traditional constructions. Because of the higher rate these construction should be preferred but an analysis if a hardware implementation is possible.

De nos jours, les liaisons par fibres optiques monomode ou multimodes équipent de plus en plus de bâtiments, de lieux et sites de transports publics (gares, aéroports, autoroutes, ports ou plateformes multimodales). Ceci est dû au débit important qu'offre la fibre optique et à son atténuation très faible sur de longues distances. Afin de gérer au mieux la transmission au sein de la fibre optique, les performances de divers codes à une dimension 1D OOC et PC et deux dimensions 2D-MWOOC sur les récepteurs conventionnels ont été étudiées. Cependant, les performances obtenues restent limitées lorsque le nombre d'utilisateurs croit significativement. En ne tenant pas compte du bruit causé par les composants du système, la dégradation des performances est principalement due aux interférences d'accès multiples (IAM). De ce fait, nous proposons dans le cadre de cette thèse d'optimiser les performances du récepteur PIC en ajoutant un facteur de compensation qui permet de réduire au mieux les interférences d'accès multiples. Ceci est étudié avec les codes optiques à deux dimensions 2D-PC/PC que nous générons. Des expérimentations sur un banc d'essai optique en laboratoire sont effectuées pour illustrer la mise en oeuvre de multiservices. Une validation de certains résultats théoriques est également menée sur un simulateur dédié.

The continuous evolution of the optical communication technology in the last few years has allowed the growing users demand for higher bandwidth to be satisfactorily attended. This new demand is mainly attributed to the growing popularity of bandwidth-intensive networking applications, such as Internet protocol television (IPTV), high-definition television (HDTV), e-learning, e-health, super high-definition (SHD) class digital movies, and e-culture based on 3-D full-HD video. These applications will eventually require differentiated service types and diversified data rates. As a result, capability of supporting differentiated-quality of service (QoS) and multirate transmission are becoming a challenge for future optical networks. Among many multiplexing techniques, optical code-division multiple-access (OCDMA) constitutes a potential candidate for next generation optical networks, particularly due to features like asynchronous operation, simplified network control, easy addition of new users, and possibility of differentiated-QoS at the physical layer. The performance of this access technique is mainly limited by multiple-access interference (MAI). Although in multirate, multiservice OCDMA systems MAI is considered to be binomially distributed, others probability distributions for the MAI analyzes and consequently the bit error rate (BER) have been used as an approximation. In this context, this dissertation proposes a new formalism to evaluate the BER performance of 1-D and 2-D multirate, multiservice OCDMA systems considering MAI as binomially distributed, resulting in a more accurate BER expression. Also, the proposed approach does not require knowledge a priori of the generated code sequences themselves, which means that the system performance analysis is achieved more easily using only the code parameters. Furthermore, a multirate OCDMA system employing 1-D optical orthogonal code (OOC) and 2-D optical fast frequency hopping (OFFH) codes is also investigated. A performance comparison in terms of BER for the OOC-based system assuming both Poisson and binomial distribution for the MAI is investigated. A further comparison of an OFFH-based multirate system assuming both gaussian and binomial distribution for the MAI is also analyzed. It is shown that BER performance in multirate scenarios can be overestimated or underestimated by many orders of magnitude depending on the assumed distribution for the MAI and on the number of simultaneous users. Moreover, it is presented for the first time a hybrid OCDM/WDM optical packet switch capable of supporting multirate and differentiated-QoS transmission. The architecture of the proposed multirate switch and its performance in terms of packet loss probability are also presented. Finally, it is shown that using a gaussian or Poisson distribution for the MAI might not be appropriate for a reliable BER estimate, since they are not acceptable approximations to assess the performance of multirate, multiservice systems with good accuracy. / A contínua evolução da tecnologia de comunicações ópticas observada nos últimos anos vem possibilitando atender a crescente demanda dos usuários por maior largura de banda. Esta nova demanda é atribuída principalmente à popularidade crescente de aplicações de banda larga, tais como TV por protocolo de Internet (IPTV), televisão de alta definição (HDTV), ensino online, e-saúde, filmes digitais de super alta definição (SHD) e e-cultura baseado em vídeo 3-D full-HD. Esta vasta gama de aplicações acabará por necessitar de tipos de serviços diferenciados e taxas de transmissão de dados diversificadas. Como resultado, capacidades de suportar diferenciamento de qualidade de serviço (QoS) e transmissão de múltiplas taxas estão se tornando um desafio para as redes ópticas futuras. Dentre muitas técnicas de multiplexação existentes, OCDMA constitui-se em um candidato potencial para as redes ópticas de próxima geração, particularmente devido a características como operação assíncrona, controle de rede simplificada, fácil adição de novos usuários, e também possibilidade de QoS diferenciado na camada física. O desempenho desta técnica de acesso múltiplo é limitado principalmente por interferência múltipla de acesso (MAI). Embora em sistemas OCDMA de múltiplas taxas e multiserviços a MAI deva ser considerada binomialmente distribuída, outras distribuições de probabilidade para a análise da MAI e, consequentemente, da BER têm sido utilizadas como uma aproximação. Neste contexto, este trabalho propõe um novo formalismo para avaliar o desempenho da BER de sistemas OCDMA 1-D e 2-D de múltiplas taxas e multisserviços considerando a MAI como binomialmente distribuída, o que resulta em uma expressão mais precisa para a BER. Além disso, o método proposto neste trabalho não requer conhecimento, a priori, das sequências de códigos geradas, o que significa que a análise do desempenho do sistema é obtida mais facilmente utilizando apenas os parâmetros do código. Um sistema OCDMA de múltiplas taxas empregando códigos 1-D OOC e 2-D OFFH será também investigado. Uma comparação do desempenho em termos de BER para o sistema baseado em OOC supondo tanto a distribuição de Poisson quanto a binomial para a MAI é investigada. De forma análoga, uma comparação para o sistema de múltiplas taxas baseado em OFFH supondo desta vez tanto a distribuição gaussiana quanto a binomial para a MAI também é analisada. É mostrado que a BER em cenários de múltiplas taxas pode ser superestimada ou subestimada em várias ordens de grandeza, dependendo da distribuição suposta para a MAI e do número de usuários simultâneos no sistema. Além disso, é apresentado pela primeira vez um switch óptico híbrido de pacotes OCDM/WDM capaz de prover transmissões de múltiplas taxas e serviços diferenciados. A arquitetura do switch proposto e seu desempenho em termos de probabilidade de perda de pacotes também são apresentados. Finalmente, é mostrado que o uso da distribuição gaussiana ou poissoniana para a MAI pode não ser apropriado para uma estimativa confiável de BER, uma vez que não são aproximações aceitáveis para avaliar o desempenho de sistemas de múltiplas taxas e multiserviços com boa precisão.

The continuous evolution of the optical communication technology in the last few years has allowed the growing users demand for higher bandwidth to be satisfactorily attended. This new demand is mainly attributed to the growing popularity of bandwidth-intensive networking applications, such as Internet protocol television (IPTV), high-definition television (HDTV), e-learning, e-health, super high-definition (SHD) class digital movies, and e-culture based on 3-D full-HD video. These applications will eventually require differentiated service types and diversified data rates. As a result, capability of supporting differentiated-quality of service (QoS) and multirate transmission are becoming a challenge for future optical networks. Among many multiplexing techniques, optical code-division multiple-access (OCDMA) constitutes a potential candidate for next generation optical networks, particularly due to features like asynchronous operation, simplified network control, easy addition of new users, and possibility of differentiated-QoS at the physical layer. The performance of this access technique is mainly limited by multiple-access interference (MAI). Although in multirate, multiservice OCDMA systems MAI is considered to be binomially distributed, others probability distributions for the MAI analyzes and consequently the bit error rate (BER) have been used as an approximation. In this context, this dissertation proposes a new formalism to evaluate the BER performance of 1-D and 2-D multirate, multiservice OCDMA systems considering MAI as binomially distributed, resulting in a more accurate BER expression. Also, the proposed approach does not require knowledge a priori of the generated code sequences themselves, which means that the system performance analysis is achieved more easily using only the code parameters. Furthermore, a multirate OCDMA system employing 1-D optical orthogonal code (OOC) and 2-D optical fast frequency hopping (OFFH) codes is also investigated. A performance comparison in terms of BER for the OOC-based system assuming both Poisson and binomial distribution for the MAI is investigated. A further comparison of an OFFH-based multirate system assuming both gaussian and binomial distribution for the MAI is also analyzed. It is shown that BER performance in multirate scenarios can be overestimated or underestimated by many orders of magnitude depending on the assumed distribution for the MAI and on the number of simultaneous users. Moreover, it is presented for the first time a hybrid OCDM/WDM optical packet switch capable of supporting multirate and differentiated-QoS transmission. The architecture of the proposed multirate switch and its performance in terms of packet loss probability are also presented. Finally, it is shown that using a gaussian or Poisson distribution for the MAI might not be appropriate for a reliable BER estimate, since they are not acceptable approximations to assess the performance of multirate, multiservice systems with good accuracy. / A contínua evolução da tecnologia de comunicações ópticas observada nos últimos anos vem possibilitando atender a crescente demanda dos usuários por maior largura de banda. Esta nova demanda é atribuída principalmente à popularidade crescente de aplicações de banda larga, tais como TV por protocolo de Internet (IPTV), televisão de alta definição (HDTV), ensino online, e-saúde, filmes digitais de super alta definição (SHD) e e-cultura baseado em vídeo 3-D full-HD. Esta vasta gama de aplicações acabará por necessitar de tipos de serviços diferenciados e taxas de transmissão de dados diversificadas. Como resultado, capacidades de suportar diferenciamento de qualidade de serviço (QoS) e transmissão de múltiplas taxas estão se tornando um desafio para as redes ópticas futuras. Dentre muitas técnicas de multiplexação existentes, OCDMA constitui-se em um candidato potencial para as redes ópticas de próxima geração, particularmente devido a características como operação assíncrona, controle de rede simplificada, fácil adição de novos usuários, e também possibilidade de QoS diferenciado na camada física. O desempenho desta técnica de acesso múltiplo é limitado principalmente por interferência múltipla de acesso (MAI). Embora em sistemas OCDMA de múltiplas taxas e multiserviços a MAI deva ser considerada binomialmente distribuída, outras distribuições de probabilidade para a análise da MAI e, consequentemente, da BER têm sido utilizadas como uma aproximação. Neste contexto, este trabalho propõe um novo formalismo para avaliar o desempenho da BER de sistemas OCDMA 1-D e 2-D de múltiplas taxas e multisserviços considerando a MAI como binomialmente distribuída, o que resulta em uma expressão mais precisa para a BER. Além disso, o método proposto neste trabalho não requer conhecimento, a priori, das sequências de códigos geradas, o que significa que a análise do desempenho do sistema é obtida mais facilmente utilizando apenas os parâmetros do código. Um sistema OCDMA de múltiplas taxas empregando códigos 1-D OOC e 2-D OFFH será também investigado. Uma comparação do desempenho em termos de BER para o sistema baseado em OOC supondo tanto a distribuição de Poisson quanto a binomial para a MAI é investigada. De forma análoga, uma comparação para o sistema de múltiplas taxas baseado em OFFH supondo desta vez tanto a distribuição gaussiana quanto a binomial para a MAI também é analisada. É mostrado que a BER em cenários de múltiplas taxas pode ser superestimada ou subestimada em várias ordens de grandeza, dependendo da distribuição suposta para a MAI e do número de usuários simultâneos no sistema. Além disso, é apresentado pela primeira vez um switch óptico híbrido de pacotes OCDM/WDM capaz de prover transmissões de múltiplas taxas e serviços diferenciados. A arquitetura do switch proposto e seu desempenho em termos de probabilidade de perda de pacotes também são apresentados. Finalmente, é mostrado que o uso da distribuição gaussiana ou poissoniana para a MAI pode não ser apropriado para uma estimativa confiável de BER, uma vez que não são aproximações aceitáveis para avaliar o desempenho de sistemas de múltiplas taxas e multiserviços com boa precisão.