Optimising the throughput performance for wireless networks is one of the challenging tasks in the objectives of communication engineering, since wireless channels are prone to errors due to path losses, random noise, and fading phenomena. The transmission errors will be worse in a multihop scenario due to its accumulative effects. Network Coding (NC) is an elegant technique to improve the throughput performance of a communication network. There is the fact that the bit error rates over one modulation symbol of 16- and higher order- Quadrature Amplitude Modulation (QAM) scheme follow a certain pattern. The Scattered Random Network Coding (SRNC) system was proposed in the literature to exploit the error pattern of 16-QAM by using bit-scattering to improve the throughput of multihop network to which is being applied the Random Linear Network Coding (RLNC). This thesis aims to improve further the SRNC system by using Forward Error Correction (FEC) code; the proposed system is called Joint RLNC and FEC with interleaving. The first proposed system (System-I) uses Convolutional Code (CC) FEC. The performances analysis of System-I with various CC rates of 1/2, 1/3, 1/4, 1/6, and 1/8 was carried out using the developed simulation tools in MATLAB and compared to two benchmark systems: SRNC system (System-II) and RLNC system (System- III). The second proposed system (System-IV) uses Reed-Solomon (RS) FEC code. Performance evaluation of System IV was carried out and compared to three systems; System-I with 1/2 CC rate, System-II, and System-III. All simulations were carried out over three possible channel environments: 1) AWGN channel, 2) a Rayleigh fading channel, and 3) a Rician fading channel, where both fading channels are in series with the AWGN channel. The simulation results show that the proposed system improves the SRNC system. How much improvement gain can be achieved depends on the FEC type used and the channel environment.

A afiliação genérica de Fischerella e Hapalosiphon é problemática devido à instabilidade dos caracteres morfológicos. Os gêneros Fischerella e Hapalosiphon são diferenciados pela presença de tricoma multisseriado e uni ou bisseriado, respectivamente. Porém, geneticamente esses caracteres não se mostraram diacríticos para diferenciar gêneros. Estudos moleculares de linhagens isoladas de ecossistemas brasileiros são escassos para Fischerella e inexistentes para Hapalosiphon. Neste estudo, oito linhagens de cianobactérias, pertencentes à família Hapalosiphonaceae, isoladas de água doce e solos brasileiros foram caracterizadas morfologicamente e geneticamente e analisadas para a produção de substâncias bioativas. As análises morfológicas identificaram cinco morfotípos de Fischerella (CENA19, CENA161, CENA212, CENA213, CENA214) e três de Hapalosiphon (CENA63, CENA71, CENA72). As análises filogenéticas do RNAr 16S usando neighbor-joining (NJ) e máxima verossimilhança (MV) colocaram todas as linhagens isoladas em um agrupamento com alto suporte (reamostragens de 99% NJ e MV) contendo membros da ordem Nostocales. Além disso, as linhagens de Fischerella selecionadas para o estudo agruparam-se em um clado interno com alto valor de reamostragem (100% NJ e 86% MV), com exceção da Fischerella CENA19. A posição dessa estirpe na árvore filogenética indica que necessita de revisão taxonômica. As linhagens de solo Hapalosiphon CENA71 e CENA72 também formaram um clado interno separado (99% NJ e 98% MV), mas a linhagem de água doce CENA63 foi colocada em um clado diferente (com valores de reamostragens de 99% NJ e MV), juntamente com linhagens do gênero Hapalosiphon e Westielopsis prolífica SAG 16.93, oriundas de solo. A comparação das análises filogenéticas individuais de regiões dos genes RNAr 16S, rpoC1, rbcL, tufA, e cpcBA-IGS das três linhagens de Hapalosiphon e de duas linhagens de Fischerella, CENA19 e CENA161, mostrou resultados incongruentes devido as diferentes taxas evolutivas desses genes. No entanto, a análise filogenética concatenada desses genes, mostrou que a Fischerella CENA19 agrupou com as duas linhagens de Hapalosiphon CENA71 e CENA72, com alto valor de reamostragem (100%), enquanto que a Fischerella CENA 161 e a Hapalosiphon CENA63 posicionaram-se cada uma em clados separados. Os resultados indicam que a nomenclatura das linhagens de cianobactérias da família Hapalosiphonaceae necessita de revisão. Os extratos intra e extracelulares das linhagens Fischerella sp. CENA161 e CENA19 e Hapalosiphon sp. CENA71 e CENA72 mostraram efeitos inibitórios no crescimento de bactérias patogênicas. As análises em espectrômetro de massas Q-TOF MS/MS indicaram a putativa presença de aeruginopeptina, cianopeptolina, fischerelina, aeruginosina, oscilapeptilida, microcistinas e ácido tumonóico nos extratos. No extrato intracelular da Fischerella sp. CENA161 identificou-se três ou quatro variantes de microcistinas, LR, LL, FR e/ou M(O)R. Fragmentos dos genes mcyA, mcyB, mcyC, mcyD, mcyE, mcyG e mcyI dessa linhagem foram seqüenciados. Nas duas análises filogenéticas realizadas com sequências de aminoácidos de McyE e sequências concatenadas de McyD, McyE e McyG, as enzimas da microcistina sintetase ficaram agrupadas de acordo com os gêneros de cianobactérias indicando um padrão de evolução / The generic affiliation of Fischerella and Hapalosiphon is problematic due to instability of morphological characters. The Fischerella and Hapalosiphon genera are differentiated by the presence of trichome multisseriate and uni or bisseriate, respectively. However, genetically these characters were not diacritical to distinguish genera. Molecular studies of strains isolated from Brazilian ecosystems are scarce for Fischerella and absent for Hapalosiphon. In this study, eight cyanobacterial strains, belonging to Hapalosiphonaceae family, isolated from Brazilian freshwater and soil were morphologically and genetically characterized and analyzed for bioactive compound productions. The morphological analyses identified five Fischerella (CENA19, CENA161, CENA212, CENA213, CENA214) and three Hapalosiphon (CENA63, CENA71, CENA72) morphotypes. The neighbor-Joining (NJ) and maximum likelihood (ML) phylogenetic analyses of 16S rRNA placed all isolated strains in high supported (99% NJ and ML of bootstrap) cluster containing members of the order Nostocales. Furthermore, the Fischerella strains studied were grouped in an internal clade with high bootstrap value (100% NJ and 86% ML), with exception of Fischerella CENA19. The position of this strain in the phylogenetic tree indicates that it needs taxonomical revision. The soil Hapalosiphon strains CENA71 and CENA72 also formed a separated tight internal clade (99% NJ and 98% ML), but the freshwater strain CENA63 was placed in a different clade (99% NJ and ML of bootstrap value) together with Hapalosiphon strains genera and Westielopsis prolifica SAG 16.93, originated from soil. The comparison of the phylogenetic analyses of individual regions of the genes 16S rRNA, rpoC1, rbcL, tufA, and cpcBA-IGS from the three Hapalosiphon strains and the two Fischerella strains CENA19 and CENA161 showed incongruent results due to different evolutionary rates of these genes. However, the concatenated phylogenetic analysis of these genes, showed that Fischerella CENA19 grouped with the two Hapalosiphon strains CENA71 and CENA72 with high bootstrap value (100%), while Fischerella CENA 161 and Hapalosiphon CENA63 were positionated each one in separate clades. The results indicate that the nomenclature of cyanobacterial strains from the family Hapalosiphonaceae needs revision. The intra and extracellular extracts of the Fischerella sp. strains CENA161 and CENA19 and Hapalosiphon sp. strains CENA71 and CENA72 showed inhibitory effects on the growth of pathogenic bacteria. The analysis in the mass spectrometer Q-TOF MS/MS indicated the presence of aeruginopeptin, cyanopeptolin, fischerellin, aeruginosin, oscillapeptilide, microcystins and tumonoic acid in the extracts. In the intracellular extracts of Fischerella sp. CENA161, three or four variants of microcystins, LR, LL, FR and/or M(O)R, were identified. Fragments of genes mcyA, mcyB, mcyC, mcyD, mcyE, mcyG and mcyI of this strain were sequenced. In both phylogenetic analyses performed with amino acid sequences of McyE and concatenated sequences of McyD, McyE and McyG, the microcystin synthetase enzymes were grouped according to the cyanobacterial genera, indicating a pattern of evolution

Optimising the throughput performance for wireless networks is one of the challenging tasks in the objectives of communication engineering, since wireless channels are prone to errors due to path losses, random noise, and fading phenomena. The transmission errors will be worse in a multihop scenario due to its accumulative effects. Network Coding (NC) is an elegant technique to improve the throughput performance of a communication network. There is the fact that the bit error rates over one modulation symbol of 16- and higher order- Quadrature Amplitude Modulation (QAM) scheme follow a certain pattern. The Scattered Random Network Coding (SRNC) system was proposed in the literature to exploit the error pattern of 16-QAM by using bit-scattering to improve the throughput of multihop network to which is being applied the Random Linear Network Coding (RLNC). This thesis aims to improve further the SRNC system by using Forward Error Correction (FEC) code; the proposed system is called Joint RLNC and FEC with interleaving. The first proposed system (System-I) uses Convolutional Code (CC) FEC. The performances analysis of System-I with various CC rates of 1/2, 1/3, 1/4, 1/6, and 1/8 was carried out using the developed simulation tools in MATLAB and compared to two benchmark systems: SRNC system (System-II) and RLNC system (System- III). The second proposed system (System-IV) uses Reed-Solomon (RS) FEC code. Performance evaluation of System IV was carried out and compared to three systems; System-I with 1/2 CC rate, System-II, and System-III. All simulations were carried out over three possible channel environments: 1) AWGN channel, 2) a Rayleigh fading channel, and 3) a Rician fading channel, where both fading channels are in series with the AWGN channel. The simulation results show that the proposed system improves the SRNC system. How much improvement gain can be achieved depends on the FEC type used and the channel environment. / Indonesian Government and the University of Bradford