O câncer de pulmão é responsável por aproximadamente 13% do total de casos de neoplasias malignas e por cerca de 1.4 milhões de mortes por ano em todo mundo. Esta neoplasia apresenta-se sob dois principais subtipos: câncer de pulmão de pequenas células (CPPC) e câncer de pulmão de não-pequenas células (CPNPC). Cerca de 85% dos casos de câncer de pulmão são do tipo CPNPC. Os sinais e sintomas são secundários ao crescimento do tumor primário, ao comprometimento lobo-regional, à disseminação à distância, ou são secundários às síndromes paraneoplásicas. Essas características refletem diretamente sobre as taxas de mortalidade; de cada 100 novos casos, 80 são inoperáveis e a maioria morre dentro de 3 anos. Isso significa que, apesar dos diversos avanços no diagnóstico e tratamento, o prognóstico do câncer de pulmão permanece sendo extremamente ruim, com sobrevida média de 10 meses, e cumulativa total em 5 anos de aproximadamente 12%. Atualmente, o prognóstico e a decisão terapêutica de pacientes com câncer de pulmão é baseada no TNM, Embora esse seja o procedimento considerado padrãoouro entre os profissionais de saúde, ele não leva em consideração características biológicas do tumor. Nesse contexto, a identificação de biomarcadores para câncer pode agregar importantes informações ao já estabelecido sistema TNM e resultar em tratamentos mais eficientes e em menores taxas de mortalidade. Existem 5 fases distintas que conceitualizam o desenvolvimento de um biomarcador tumoral. Através dessas fases consecutivas, é possível que se desenvolvam ferramentas úteis para triagem populacional, capazes de serem implementadas na rotina clínica para predição de desfecho do paciente, resposta terapêutica e monitoramento da doença. O presente projeto avaliou o valor prognóstico dos principais genes citados na literatura como potenciais biomarcadores para CPNPC, e verificou-se que nenhum deles apresentou significância na correlação estatística que indica poder prognóstico. Além disso, identificamos e validamos o papel prognóstico da cofilina-1 por meio de dados de microarranjo e quantificação de seu imunoconteúdo em biópsias de CPNPC. Para tanto, fizemos uso de meta-análise de bancos de dados e análise densitométrica das reações imuno-histoquímicas, seguida de correlação com dados de grau de diferenciação tumoral, classificação histológica, sexo, idade e desfecho relativo a cada caso. Além disso, desenvolvemos um método de baixo custo, fácil execução e ampla aplicação e reproducibilidade, capaz de quantificar a proteína em amostras biológicas, com potencial para ser implementado na rotina clínica e aplicamos esse método em uma coorte restrospectiva de CPNPC. Confirmamos assim o papel prognóstico da cofilina-1. Estes achados seguem a lógica das fases de desenvolvimento de um biomarcador e representam um grande passo no seu processo de validação. / Lung cancer accounts for approximately 13% of all malignant tumor cases and for about 1.4 million deaths per year worldwide. This cancer has two main subtypes: Small Cell Lung Cancer (SCLC) and Non-Small Cell Lung Cancer (NSCLC). About 85% of cases of lung cancer are NSCLC type. The signs and symptoms are secondary to the primary tumor growth, to regional lobe commitment and distant spread, or are secondary to paraneoplastic syndromes. These features reflect directly on mortality rates; 80 in every 100 new cases are inoperable and most die within 3 years. This means that, despite many advances in diagnosis treatment, the prognosis of lung cancer remains extremely poor, with median survival of 10 months, and total cumulative survival in 5-year of approximately 12%. Currently, prognosis and therapeutic decisions in patients with lung cancer is based on TNM. Although this procedure is considered gold standard among health professionals, it does not take into account the biological characteristics of the tumor. In this context, the identification of cancer biomarkers may add important information to the already established TNM system and result in better treatments and lower mortality rates. There are five distinct phases that conceptualize a tumor biomarker development of. Through these successive phases, it is possible to develop useful tools for population screening, capable of implementation in clinical practice for prediction of patient outcome, therapeutic response and disease monitoring. This project evaluated the prognostic value of major genes mentioned in literature as potential biomarkers for NSCLC and found that none of them showed statistical significance in the correlation that indicates prognostic power. It also identified and validated the prognostic role of cofilin-1 by microarray data and quantification of their immunocontent in biopsies of NSCLC. For this purpose, we used data metaanalysis and immunohistochemical reactions densitometric analysis, followed by correlation with data from tumor grade, histological classification, sex, age and outcome for each case. In addition, we developed a low-cost protocol, of easy implementation and wide application and reproducibility, able to quantify the protein in biological samples, with the potential to be implemented in clinical practice. We applied this method in a retrospective cohort of NSCLC and confirm the prognostic role of cofilin-1. These findings follow the logical phases of biomarker development and represent a major step in its validation process.

The exponentially increasing demand for high-speed wireless communications will no longer be satisfied by the traditional radio frequency (RF) in the near future due to its limited spectrum and overutilization. To resolve this imminent issue, industrial and research communities have been looking into alternative technologies for communication. Among them, visible light communication (VLC) has attracted much attention because it utilizes the unlicensed, free and safe spectrum, whose bandwidth is thousand times larger than the entire RF spectrum. Moreover, VLC can be integrated into existing lighting systems to offer a dual-purpose, cost-effective and energy-efficient solution for next-generation small-cell networks (SCNs), giving birth to the concept of optical attocell networks. Most relevant works in the literature rely on system simulations to quantify the performance of attocell networks, which suffer from high computational complexity and provide limited insights about the network. Mathematical tools, on the other hand, are more tractable and scalable and are shown to closely approximate practical systems. The presented work utilizes stochastic geometry for downlink evaluation of optical attocell networks, where the co-channel interference (CCI) surpasses noise and becomes the limiting factor of the link throughput. By studying the moment generating function (MGF) of the aggregate interference, a theoretical framework for modeling the distribution of signal-to-interference-plus-noise ratio (SINR) is presented, which allows important performance metrics such as the coverage probability and link throughput to be derived. Depending on the source of interference, CCI can be classified into two categories: inter-cell interference (ICI) and intra-cell interference. In this work, both types of interference are characterized, based on which effective interference mitigation techniques such as the coordinated multipoint (CoMP), power-domain multiplexing and successive interference cancellation (SIC) are devised. The proposed mathematical framework is applicable to attocell networks with and without such interference mitigation techniques. Compared to RF networks, optical attocell networks are inherently more secure in the physical layer because visible light does not penetrate through opaque walls. This work analytically quantifies the physical-layer security of attocell networks from an information-theoretic point of view. Secrecy enhancement techniques such as AP cooperation and eavesdropper-free protected zones are also discussed. It is shown that compared to AP cooperation, implementing secrecy protected zones is more effective and it can contribute significantly to the network security.

O câncer de pulmão é responsável por aproximadamente 13% do total de casos de neoplasias malignas e por cerca de 1.4 milhões de mortes por ano em todo mundo. Esta neoplasia apresenta-se sob dois principais subtipos: câncer de pulmão de pequenas células (CPPC) e câncer de pulmão de não-pequenas células (CPNPC). Cerca de 85% dos casos de câncer de pulmão são do tipo CPNPC. Os sinais e sintomas são secundários ao crescimento do tumor primário, ao comprometimento lobo-regional, à disseminação à distância, ou são secundários às síndromes paraneoplásicas. Essas características refletem diretamente sobre as taxas de mortalidade; de cada 100 novos casos, 80 são inoperáveis e a maioria morre dentro de 3 anos. Isso significa que, apesar dos diversos avanços no diagnóstico e tratamento, o prognóstico do câncer de pulmão permanece sendo extremamente ruim, com sobrevida média de 10 meses, e cumulativa total em 5 anos de aproximadamente 12%. Atualmente, o prognóstico e a decisão terapêutica de pacientes com câncer de pulmão é baseada no TNM, Embora esse seja o procedimento considerado padrãoouro entre os profissionais de saúde, ele não leva em consideração características biológicas do tumor. Nesse contexto, a identificação de biomarcadores para câncer pode agregar importantes informações ao já estabelecido sistema TNM e resultar em tratamentos mais eficientes e em menores taxas de mortalidade. Existem 5 fases distintas que conceitualizam o desenvolvimento de um biomarcador tumoral. Através dessas fases consecutivas, é possível que se desenvolvam ferramentas úteis para triagem populacional, capazes de serem implementadas na rotina clínica para predição de desfecho do paciente, resposta terapêutica e monitoramento da doença. O presente projeto avaliou o valor prognóstico dos principais genes citados na literatura como potenciais biomarcadores para CPNPC, e verificou-se que nenhum deles apresentou significância na correlação estatística que indica poder prognóstico. Além disso, identificamos e validamos o papel prognóstico da cofilina-1 por meio de dados de microarranjo e quantificação de seu imunoconteúdo em biópsias de CPNPC. Para tanto, fizemos uso de meta-análise de bancos de dados e análise densitométrica das reações imuno-histoquímicas, seguida de correlação com dados de grau de diferenciação tumoral, classificação histológica, sexo, idade e desfecho relativo a cada caso. Além disso, desenvolvemos um método de baixo custo, fácil execução e ampla aplicação e reproducibilidade, capaz de quantificar a proteína em amostras biológicas, com potencial para ser implementado na rotina clínica e aplicamos esse método em uma coorte restrospectiva de CPNPC. Confirmamos assim o papel prognóstico da cofilina-1. Estes achados seguem a lógica das fases de desenvolvimento de um biomarcador e representam um grande passo no seu processo de validação. / Lung cancer accounts for approximately 13% of all malignant tumor cases and for about 1.4 million deaths per year worldwide. This cancer has two main subtypes: Small Cell Lung Cancer (SCLC) and Non-Small Cell Lung Cancer (NSCLC). About 85% of cases of lung cancer are NSCLC type. The signs and symptoms are secondary to the primary tumor growth, to regional lobe commitment and distant spread, or are secondary to paraneoplastic syndromes. These features reflect directly on mortality rates; 80 in every 100 new cases are inoperable and most die within 3 years. This means that, despite many advances in diagnosis treatment, the prognosis of lung cancer remains extremely poor, with median survival of 10 months, and total cumulative survival in 5-year of approximately 12%. Currently, prognosis and therapeutic decisions in patients with lung cancer is based on TNM. Although this procedure is considered gold standard among health professionals, it does not take into account the biological characteristics of the tumor. In this context, the identification of cancer biomarkers may add important information to the already established TNM system and result in better treatments and lower mortality rates. There are five distinct phases that conceptualize a tumor biomarker development of. Through these successive phases, it is possible to develop useful tools for population screening, capable of implementation in clinical practice for prediction of patient outcome, therapeutic response and disease monitoring. This project evaluated the prognostic value of major genes mentioned in literature as potential biomarkers for NSCLC and found that none of them showed statistical significance in the correlation that indicates prognostic power. It also identified and validated the prognostic role of cofilin-1 by microarray data and quantification of their immunocontent in biopsies of NSCLC. For this purpose, we used data metaanalysis and immunohistochemical reactions densitometric analysis, followed by correlation with data from tumor grade, histological classification, sex, age and outcome for each case. In addition, we developed a low-cost protocol, of easy implementation and wide application and reproducibility, able to quantify the protein in biological samples, with the potential to be implemented in clinical practice. We applied this method in a retrospective cohort of NSCLC and confirm the prognostic role of cofilin-1. These findings follow the logical phases of biomarker development and represent a major step in its validation process.

abstract: Small Cell Carcinoma of the Ovary Hypercalcemic Type (SCCOHT) is a rare and highly aggressive ovarian cancer that affects children and young women at a mean age of 24 years. Most SCCOHT patients are diagnosed at an advanced stage and do not respond to chemotherapy. As a result, more than 75% of patients succumb to their disease within 1-2 years. To provide insights into the biological, diagnostic, and therapeutic vulnerabilities of this deadly cancer, a comprehensive characterization of 22 SCCOHT cases and 2 SCCOHT cell lines using microarray and next-generation sequencing technologies was performed. Following histological examination, tumor DNA and RNA were extracted and used for array comparative genomic hybridization and gene expression microarray analyses. In agreement with previous reports, SCCOHT presented consistently diploid profiles with few copy number aberrations. Gene expression analysis showed SCCOHT tumors have a unique gene expression profile unlike that of most common epithelial ovarian carcinomas. Dysregulated cell cycle control, DNA repair, DNA damage-response, nucleosome assembly, neurogenesis and nervous system development were all characteristic of SCCOHT tumors. Sequencing of DNA from SCCOHT patients and cell lines revealed germline and somatic inactivating mutations in the SWI/SNF chromatin-remodeling gene SMARCA4 in 79% (19/24) of SCCOHT patients in addition to SMARCA4 protein loss in 84% (16/19) of SCCOHT tumors, but in only 0.4% (2/485) of other primary ovarian tumors. Ongoing studies are now focusing on identifying treatments for SCCOHT based on therapeutic vulnerabilities conferred by ubiquitous inactivating mutations in SMARCA4 in addition to gene and protein expression data. Our characterization of the molecular landscape of SCCOHT and the breakthrough identification of inactivating SMARCA4 mutations in almost all cases of SCCOHT offers the first significant insight into the molecular pathogenesis of this disease. The loss of SMARCA4 protein is a highly sensitive and specific marker of the disease, highlighting its potential role as a diagnostic marker, and offers the opportunity for genetic testing of family members at risk. Outstanding questions remain about the role of SMARCA4 loss in the biology, histogenesis, diagnosis, and treatment of SCCOHT. / Dissertation/Thesis / Doctoral Dissertation Molecular and Cellular Biology 2014

O câncer de pulmão é responsável por aproximadamente 13% do total de casos de neoplasias malignas e por cerca de 1.4 milhões de mortes por ano em todo mundo. Esta neoplasia apresenta-se sob dois principais subtipos: câncer de pulmão de pequenas células (CPPC) e câncer de pulmão de não-pequenas células (CPNPC). Cerca de 85% dos casos de câncer de pulmão são do tipo CPNPC. Os sinais e sintomas são secundários ao crescimento do tumor primário, ao comprometimento lobo-regional, à disseminação à distância, ou são secundários às síndromes paraneoplásicas. Essas características refletem diretamente sobre as taxas de mortalidade; de cada 100 novos casos, 80 são inoperáveis e a maioria morre dentro de 3 anos. Isso significa que, apesar dos diversos avanços no diagnóstico e tratamento, o prognóstico do câncer de pulmão permanece sendo extremamente ruim, com sobrevida média de 10 meses, e cumulativa total em 5 anos de aproximadamente 12%. Atualmente, o prognóstico e a decisão terapêutica de pacientes com câncer de pulmão é baseada no TNM, Embora esse seja o procedimento considerado padrãoouro entre os profissionais de saúde, ele não leva em consideração características biológicas do tumor. Nesse contexto, a identificação de biomarcadores para câncer pode agregar importantes informações ao já estabelecido sistema TNM e resultar em tratamentos mais eficientes e em menores taxas de mortalidade. Existem 5 fases distintas que conceitualizam o desenvolvimento de um biomarcador tumoral. Através dessas fases consecutivas, é possível que se desenvolvam ferramentas úteis para triagem populacional, capazes de serem implementadas na rotina clínica para predição de desfecho do paciente, resposta terapêutica e monitoramento da doença. O presente projeto avaliou o valor prognóstico dos principais genes citados na literatura como potenciais biomarcadores para CPNPC, e verificou-se que nenhum deles apresentou significância na correlação estatística que indica poder prognóstico. Além disso, identificamos e validamos o papel prognóstico da cofilina-1 por meio de dados de microarranjo e quantificação de seu imunoconteúdo em biópsias de CPNPC. Para tanto, fizemos uso de meta-análise de bancos de dados e análise densitométrica das reações imuno-histoquímicas, seguida de correlação com dados de grau de diferenciação tumoral, classificação histológica, sexo, idade e desfecho relativo a cada caso. Além disso, desenvolvemos um método de baixo custo, fácil execução e ampla aplicação e reproducibilidade, capaz de quantificar a proteína em amostras biológicas, com potencial para ser implementado na rotina clínica e aplicamos esse método em uma coorte restrospectiva de CPNPC. Confirmamos assim o papel prognóstico da cofilina-1. Estes achados seguem a lógica das fases de desenvolvimento de um biomarcador e representam um grande passo no seu processo de validação. / Lung cancer accounts for approximately 13% of all malignant tumor cases and for about 1.4 million deaths per year worldwide. This cancer has two main subtypes: Small Cell Lung Cancer (SCLC) and Non-Small Cell Lung Cancer (NSCLC). About 85% of cases of lung cancer are NSCLC type. The signs and symptoms are secondary to the primary tumor growth, to regional lobe commitment and distant spread, or are secondary to paraneoplastic syndromes. These features reflect directly on mortality rates; 80 in every 100 new cases are inoperable and most die within 3 years. This means that, despite many advances in diagnosis treatment, the prognosis of lung cancer remains extremely poor, with median survival of 10 months, and total cumulative survival in 5-year of approximately 12%. Currently, prognosis and therapeutic decisions in patients with lung cancer is based on TNM. Although this procedure is considered gold standard among health professionals, it does not take into account the biological characteristics of the tumor. In this context, the identification of cancer biomarkers may add important information to the already established TNM system and result in better treatments and lower mortality rates. There are five distinct phases that conceptualize a tumor biomarker development of. Through these successive phases, it is possible to develop useful tools for population screening, capable of implementation in clinical practice for prediction of patient outcome, therapeutic response and disease monitoring. This project evaluated the prognostic value of major genes mentioned in literature as potential biomarkers for NSCLC and found that none of them showed statistical significance in the correlation that indicates prognostic power. It also identified and validated the prognostic role of cofilin-1 by microarray data and quantification of their immunocontent in biopsies of NSCLC. For this purpose, we used data metaanalysis and immunohistochemical reactions densitometric analysis, followed by correlation with data from tumor grade, histological classification, sex, age and outcome for each case. In addition, we developed a low-cost protocol, of easy implementation and wide application and reproducibility, able to quantify the protein in biological samples, with the potential to be implemented in clinical practice. We applied this method in a retrospective cohort of NSCLC and confirm the prognostic role of cofilin-1. These findings follow the logical phases of biomarker development and represent a major step in its validation process.

Submitted by Edineia Teixeira (edineia.teixeira@unioeste.br) on 2018-02-23T18:26:15Z No. of bitstreams: 2 Jose_santos2017.pdf: 2362954 bytes, checksum: 3cd8da60eb3efb9e2ed240ef51fa9a23 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-02-23T18:26:15Z (GMT). No. of bitstreams: 2 Jose_santos2017.pdf: 2362954 bytes, checksum: 3cd8da60eb3efb9e2ed240ef51fa9a23 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-08-01 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Lung cancer is highly lethal and smoking is an important risk factor. Non-small cell tumor is the most common, less aggressive type of lung carcinoma and treatment involves surgery, chemotherapy, and radiation therapy. However, late diagnosis, due to the absence of signs and symptoms in the early stages of the disease, makes the survival rate low. In this sense, the search for new chemical substances with characteristics of selectivity, effectiveness and low toxicity have been investigated in the treatment of cancer. Studies show that sericin, a protein extracted from silkworm cocoons, exhibits anti-tumor and anticarcinogenic activity in colon and skin cancer cells, stimulating apoptosis and disrupting the cell cycle, and protecting normal cells against oxidative stress and lipid peroxidation . Thus, the therapeutic potential of sericin has raised interest in evaluating its effect on non-small cell lung cancer cell line. The studies were conducted in culture, where the effects of silk protein were evaluated: on cell viability, by neutral red assays and tetrazolium - MTT cytotoxicity; in the apoptotic potential, with annexin-5 and Alexa Fluor® and Propidium Iodide assays; and in cell migration by the Wound Healing model assay. Low doses of sericin were able to increase lysosomal viability, reduce mitochondrial viability, increase apoptosis and cell migration, while high doses of sericin exponentially reduced cell migration and did not alter the rate of apoptosis / necrosis of cancer cells. Sericin is a biomaterial that causes biological effects on the cell line tested, and can be used to increase lysosomal viability, reduce mitochondrial function, increase apoptosis at low doses and inhibit high-dose cell migration. / O câncer de pulmão apresenta alta letalidade e o fumo se constitui como um importante fator de risco. O tumor de células não pequenas é o tipo de carcinoma pulmonar mais comum, menos agressivo e o tratamento envolve cirurgia, quimioterapia e radioterapia. Entretanto, o diagnóstico tardio, em função da ausência de sinais e sintomas nos estágios iniciais da doença, faz com que a taxa de sobrevivência seja baixa. Nesse sentido, a busca de novas substâncias químicas com características de seletividade, de efetividade e de baixa toxicidade tem se refletido em pesquisas que investigam o uso dessas substâncias no tratamento do câncer. Estudos mostram que a sericina, proteína extraída dos casulos do bicho-da-seda, apresenta atividade antitumoral e anticarcinogênica em células cancerosas de cólon e de pele, estimulando a apoptose e interrompendo o ciclo celular, além de proteger células normais contra estresse oxidativo e peroxidação lipídica. Assim, o potencial terapêutico da sericina suscitou o interesse em avaliar seu efeito em linhagem celular de câncer de pulmão do tipo células não pequenas. Os estudos foram conduzidos em cultura, em que foram avaliados os efeitos da proteína da seda: na viabilidade celular, por meio de ensaios pelo vermelho neutro e de citotoxicidade com o tetrazólio - MTT; no potencial apoptótico, com ensaios com anexina-5 e Alexa Fluor® e Iodeto de Propídio; e na migração celular, ensaio com o modelo Wound Healing. Baixas doses de sericina foram capazes de aumentar a viabilidade lisossomal, reduzir a viabilidade mitocondrial, aumentar a apoptose e a migração celular, enquanto altas doses de sericina reduziram exponencialmente a migração celular e não alteraram a taxa de apoptose/necrose das células cancerosas. A sericina é um biomaterial que provoca efeitos biológicos na linhagem celular testada, e pode ser utilizada para aumentar a viabilidade lisossomal, reduzir a função mitocondrial, aumentar a apoptose em baixas doses e inibir a migração celular em doses altas.

Abstract The aim of this thesis is to devise novel cooperative paradigms that allow small base stations (SBSs) to perform joint optimizations in the field of interference and spectral resource management, in an automated, self-organizing way. Fostering cooperation requires a careful negotiation process and incurs additional operational costs that can ultimately limit the scalability of such an approach. Hence, we design cooperative models around the self-organizing functions of small cell networks, through which each SBS dynamically adapts its own cooperative strategy to the current network status. In the first part of the thesis, we study the co-tier interference in the downlink of an underlay small cell tier. Showing that mutual interference is a limiting factor, we propose a cooperative scheme based on the concept of interference alignment. Thereby, the SBSs autonomously devise their cooperative strategy, select partner small base stations and suppress the mutual interference via alignment. In the second part, we extend the above problem by considering additional constraints posed by non-cooperative macro cell users sharing the spectrum with the SBSs. Due to the dimension of such a problem, we show that interference alignment solutions do not often exist, hence, we propose a distributed solution based on the concept of interference draining for solving the interfering signals in the space-frequency domain. Finally, we focus on inter-tier cooperation and propose an adaptive hybrid access policy based on spectrum leasing, which jointly enhances the SBSs' transmission capacity and the macro cell users' QoS . In the studied scenarios, we highlight how SBSs' individual performance is network-dependent, due to the shared nature of the spectrum and the network infrastructure. For addressing such issues, we use novel concepts from coalitional games in partition form in which the strategic decisions are optimized by accounting for external effects, such as interference or dynamic spectrum allocation. Based on the properties of optimality of coalitional games in partition form, the proposed cooperative solutions show two key features. First, by leveraging on decentralized strategic decisions, complex interference management techniques, such as interference alignment or draining, can be implemented in practical, scalable way. Second, although notable cooperative solutions exists for certain problem dimensions, the proposed algorithms are shown to be effective for a wide range of network sizes, by achieving significant data rate enhancement and low overhead. / Tiivistelmä Työssä käsitellään merkittäviä häiriön ja radioresurssien hallintaongelmia ja ehdotetaan käytännöllisiä yhteistoiminnallisuusmalleja, jotka tuottavat merkittäviä suorituskykyparannuksia, ja samaan aikaan ottavat huomioon yhteistyön kustannukset ja strategisten päätösten vaikutuksen piensolujen keskinäiseen suorituskykyyn. Työn ensimmäisessa osassa tutkitaan makro- ja piensoluverkon välistä häiriötä laskevalla siirtosuunnalla. Koska keskinäishäiriö osoittautuu merkittäväksi rajoittavaksi tekijäksi, ehdotetaan yhteistyömenetelmää, joka perustuu ns. häiriön laskostamiseen. Tällaisessa menetelmässä pientukiasemat päättävät itsenäisesti yhteistyöstrategiansa ja valitsevat ne tukiasemat, joiden kanssa tehdään yhteistyötä ja vaimennetaan keskinäishäiriötä laskostamalla. Työn toisessa osassa edellämainittua ongelmaa laajennataan ottamalla huomioon rajoitteet, jotka aiheutuvat samoja taajuusresursseja jakavista, mutta piensolutukiasemien kanssa yhteistyötä tekemättömistä makrosoluista. Ongelman laajuudesta havaitaan, että yleensä häiriön laskostusratkaisua ei löydy tällaisessa ongelmassa. Niinpä ehdotetaan hajautettua häiriön kohdennukseen perustuvaa ratkaisua, jossa häiriösignaalit lomitellaan tila-taajuustasossa. Työn viimeisessä osassa keskitytään eri tyyppisten tukiasemasolujen kesken käytävään yhteistyöhön ja ehdotetaan mukautuvaa taajuusspektrin vuokraamiseen perustuvaa hybridi-käyttöoikeusmenetelmää, joka parantaa piensolujen kapasiteettia ja makrosolukäyttäjien palvelun laatua. Kaikissa tutkituissa menetelmissä korostetaan kuinka yksittäisen pientukiaseman suorituskyky on verkosta riippuvainen johtuen jaetusta taajuuskaistasta ja verkkoinfrastuktuurista. Näissä skenaarioissa käytetään uusia ositukseen perustuvia ns. liittoumapelikonsepteja. Tämän tyyppisissä peleissä strategiset päätökset on optimoitu ottamalla huomioon ulkoiset tekijät kuten häiriö ja dynaaminen taajuuden allokointi. Ositukseen perustuvien liitoumapelien optimaalisuus- ja stabiilisuusominaisuuksien pohjalta ehdotetaan hajautettuja algoritmeja tutkittuihin ongelmiin. Ehdotetut yhteistoiminnalliset ratkaisut osoittavat kaksi avaintekijää. Ensinnäkin käyttämällä hajautettuja strategisia päätöksiä kompleksiset häiriönhallintatekniikat, kuten häiriön lomittelu ja kohdennus, voidaan toteuttaa käytännöllisesti siten, että hyöty-kustannussuhde on tasapainossa. Toiseksi, vaikka merkittäviä yhteistyöratkaisuja on olemassa tietyn dimensioisiin ongelmiin, ehdotetut algoritmit ovat osoittautuneet tehokkaiksi hyvin erikokoisissa verkoissa. Samalla on saavutettu merkittäviä datanopeusparannuksia ja hyvä skaalautuvuus.

Récemment, il y a eu une hausse massive du trafic dans les réseaux mobiles à cause de nouveaux services et applications. Les architectures actuelles des réseaux cellulaires ne sont plus capables de gérer de façon satisfaisante ce trafic. Les Réseaux de Petites Cellules (RPC), basées sur un déploiement dense de stations de bases portables, autoorganisantes et efficaces en termes d’énergie apparait comme une solution prometteuse à ce problème. Les RPC augmentent la capacité du réseau, réduisent sa consommation énergétique et améliorent sa couverture. Par contre, elles posent des défis importants en termes de design optimal. Dans cette thèse, des aspects liés au design cellulaire et à l’allocation de ressources dans les RPC sont traités. La thèse se compose de deux parties. Dans la première partie, le design cellulaire est étudié : une population statique d’utilisateurs est considérée, et la taille optimale de cellule maximisant le débit spatial est donnée en fonction du modèle de récepteur, des conditions radio et des partitions indoor/outdoor. En considérant des utilisateurs mobiles, la taille de cellule optimale est étudiée afin de minimiser le temps de service, et minimiser le blocage et la déconnexion en cours de communication, en fonction de la vitesse des utilisateurs et du type de trafic. Le problème de placement des stations de base optimal est traité en fonction de différents critères de qualité (maximisation de débit total, équité proportionnelle, minimisation de délai, équité max-min) pour différentes distributions d’utilisateurs et partitions de cellules. Le problème de scaling de capacité dans un RPC limité par l’interférence avec pré-codage est étudié, et la quantité optimale d’antennes par utilisateurs en fonction de l’interférence inter-cellules est dérivée. Dans le cadre d’un réseau “green”, pour une charge du réseau donnée, on étudie les politiques optimales en boucle ouverte, afin de maximiser soit une fonction coût du système (contrôle centralisé) soit des fonctions de coût de chacune des stations de base (contrôle distribué). Dans la seconde partie, nous étudions l’allocation de ressources, nous introduisons les concepts de d’équité T-échelle et équité multi-échelle. Ces concepts permettent de distribuer les ressources équitablement pour les différentes classes de trafic. Ces concepts sont illustrés par des applications au partage de spectre et à l’allocation de ressources dans les femto-cellules indoor/outdoor. L’allocation de puissance pour satisfaire les demandes de trafic des utilisateurs avec un grand nombre d’interféreurs est une tâche difficile. Ce problème est abordé, et nous proposons un algorithme universel qui converge vers une configuration de puissance optimale qui satisfait les demandes des utilisateurs dans toutes les stations de base. Les performances de l’algorithme sont illustrées pour différentes configurations du système et différents niveaux de coopération entre les stations de base. / An ever increasing demand for mobile broadband applications and services is leading to a massive network densification. The current cellular system architectures are both economically and ecologically limited to handle this. The concept of small-cell networks (SCNs) based on the idea of dense deployment of self-organizing; low-cost, low-power base station (BSs) is a promising alternative. Although SCNs have the potential to significantly increase the capacity and coverage of cellular networks while reducing their energy consumption, they pose many new challenges to the optimal system design. Due to small cell sizes, the mobile users cross over many cells during the course of their service resulting in frequent handovers. Also, due to proximity of BSs, users (especially those at cell edges) experience a higher degree of interference from neighboring BSs. If one has to derive advantages from SCNs, these alleviated effects have to be taken care either by compromising on some aspects of optimality (like dedicating extra resources) or by innovating smarter algorithms or by a combination of the two. The concept of umbrella cells is introduced to take care of frequent handovers. Here extra resources are dedicated to ensure that the calls are not dropped within an umbrella cell. To manage interference, one might have to ensure that the neighboring cells always operate in independent channels or design algorithms which work well in interference dominant scenarios or use the backhaul to incorporate BS cooperation techniques. Further, small cell BS are most often battery operated, which calls for efficient power utilization and energy conservation techniques. Also, when deployed in urban areas, some of the small cells can have larger concentration of users throughout the cell, for example, hot-spots, which call in for design of SCNs with dense users. Also, with portable BSs, one has the choice to install them on street infrastructure or within residential complexes. In such cases, cell design and resource allocation has to consider aspects like user density, distribution (indoor/outdoor), mobility, attenuation, etc. We present the thesis in two parts. In the first part we study the cell design aspects, while the second part deals with the resource allocation. While the focus is on SCNs, some of the results derived and the tools and techniques used are also applicable to conventional cellular systems.

No description available.

京都大学 / 新制・課程博士 / 博士(人間健康科学) / 甲第23121号 / 人健博第83号 / 新制||人健||6(附属図書館) / 京都大学大学院医学研究科人間健康科学系専攻 / (主査)教授 椎名 毅, 教授 藤井 康友, 教授 平井 豊博 / 学位規則第4条第1項該当 / Doctor of Human Health Sciences / Kyoto University / DFAM