Vehicular networking is envisioned to be a key technology area for significant growth in the coming years. Although the expectations for this emerging technology are set very high, many practical aspects remain still unsolved for a vast deployment of vehicular networks. This dissertation addresses the enabling physical layer techniques to meet the challenges in vehicular networks operating in mobile wireless environments. Considering the infrastructure-less nature of vehicular networks, we envision cooperative diversity well positioned to meet the demanding requirements of vehicular networks with their underlying distributed structure. Cooperative diversity has been proposed as a powerful means to enhance the performance of high-rate communications over wireless fading channels. It realizes spatial diversity advantages in a distributed manner where a node uses others antennas to relay its message creating a virtual antenna array. Although cooperative diversity has garnered much attention recently, it has not yet been fully explored in the context of vehicular networks considering the unique characteristics of vehicular networks, this dissertation provides an error performance analysis study of cooperative transmission schemes for various deployment and traffic scenarios. In the first part of this dissertation, we investigate the performance of a cooperative vehicle-to-vehicle (V2V) system with amplify-and-forward relaying for typical traffic scenarios under city/urban settings and a highway area. We derive pairwise error probability (PEP) expressions and demonstrate the achievable diversity gains. The effect of imperfect channel state information (CSI) is also studied through an asymptotical PEP analysis. We present Monte-Carlo simulations to confirm the analytical derivations and present the error rate performance of the vehicular scheme with perfect and imperfect-CSI. In the second part, we consider road-to-vehicle (R2V) communications in which roadside access points use cooperating vehicles as relaying terminals. Under the assumption of decode-and-forward relaying, we derive PEP expressions for single-relay and multi-relay scenarios. In the third part, we consider a cooperative multi-hop V2V system in which direct transmission is not possible and investigate its performance through the PEP derivation and diversity gain analysis. Monte-Carlo simulations are further provided to con firm the analytical derivations and provide insight into the error rate performance improvement.

Vehicular networking is envisioned to be a key technology area for significant growth in the coming years. Although the expectations for this emerging technology are set very high, many practical aspects remain still unsolved for a vast deployment of vehicular networks. This dissertation addresses the enabling physical layer techniques to meet the challenges in vehicular networks operating in mobile wireless environments. Considering the infrastructure-less nature of vehicular networks, we envision cooperative diversity well positioned to meet the demanding requirements of vehicular networks with their underlying distributed structure. Cooperative diversity has been proposed as a powerful means to enhance the performance of high-rate communications over wireless fading channels. It realizes spatial diversity advantages in a distributed manner where a node uses others antennas to relay its message creating a virtual antenna array. Although cooperative diversity has garnered much attention recently, it has not yet been fully explored in the context of vehicular networks considering the unique characteristics of vehicular networks, this dissertation provides an error performance analysis study of cooperative transmission schemes for various deployment and traffic scenarios. In the first part of this dissertation, we investigate the performance of a cooperative vehicle-to-vehicle (V2V) system with amplify-and-forward relaying for typical traffic scenarios under city/urban settings and a highway area. We derive pairwise error probability (PEP) expressions and demonstrate the achievable diversity gains. The effect of imperfect channel state information (CSI) is also studied through an asymptotical PEP analysis. We present Monte-Carlo simulations to confirm the analytical derivations and present the error rate performance of the vehicular scheme with perfect and imperfect-CSI. In the second part, we consider road-to-vehicle (R2V) communications in which roadside access points use cooperating vehicles as relaying terminals. Under the assumption of decode-and-forward relaying, we derive PEP expressions for single-relay and multi-relay scenarios. In the third part, we consider a cooperative multi-hop V2V system in which direct transmission is not possible and investigate its performance through the PEP derivation and diversity gain analysis. Monte-Carlo simulations are further provided to con firm the analytical derivations and provide insight into the error rate performance improvement.

Ferramentas computacionais, baseadas em modelos de inventários, que consigam predizer, com precisão apropriada, o consumo de combustíveis e as emissões veiculares, principais motivadores do desenvolvimento automotivo atual, são fundamentais para a elaboração de políticas públicas eficazes vinculadas a essas questões. No Brasil, apesar de afetarem significativamente a sociedade, muitas das intervenções governamentais nesse segmento são realizadas sem a adequada avaliação de seus impactos. Isso ocorre ou porque a importância dessas ferramentas nem sempre é reconhecida ou por não se ter ferramentas apropriadas. O objetivo deste trabalho foi o de desenvolver uma ferramenta de prognóstico do consumo de combustíveis e das emissões da frota rodoviária, estatisticamente consistente, que pudesse ser utilizada para esse propósito. Para tanto, inicialmente buscou-se identificar os aspectos relacionados às questões que deveriam ser considerados nesse desenvolvimento, por meio do levantamento das tendências evolutivas e alternativas que estão sendo apresentadas relativas a combustíveis, tecnologia veicular e sistemas de transportes. A seguir foram revistos os principais modelos e ferramentas públicos, nacionais e estrangeiros, dessa natureza. E, diante da constatação que eles não poderiam ser empregados, apesar dos limites impostos pela disponibilidade de dados, foi proposta uma nova abordagem para se atingir esse objetivo. Primeiro foi desenvolvido um conjunto de planilhas integrando todos os dados e cálculos de um modelo bottom-up similar ao utilizado nos inventários de emissões tóxicas nacionais, totalmente interconectado e configurado para facilitar, por meio de um processo iterativo, o ajuste fino das estimativas mais incertas, de forma que os consumos totalizados resultantes do modelo coincidissem tanto quanto possível com os consumos observados no país. A seguir, foram desenvolvidos modelos econométricos, para estimar os consumos totalizados de combustíveis a partir de indicadores econômicos pertinentes. E pela inserção desses modelos econométricos no modelo bottom-up anterior, foi gerado um modelo híbrido que permite inventariar e prognosticar, segregadas por classes, o consumo de combustíveis e as emissões veiculares. Os consumos de combustíveis obtidos por meio desses modelos, quando comparados com valores observados, forneceram resultados estatisticamente robustos, que podem ser, em algumas condições, convertidos na emissão de CO2. O mesmo não é possível afirmar com relação às demais emissões, inclusive em função da dificuldade de vincular as fontes emissoras a indicadores de qualidade do ar, o que não invalida o uso do modelo híbrido para obtenção de resultados comparativos. Finalmente, os modelos foram submetidos a análise sensibilidade e sua aplicabilidade foi verificada para alguns cenários. / Computational tools, based on inventory models, which are able to predict, with the appropriated accuracy, vehicular fuel consumption and emissions, main current drivers of the automotive development, are essentials for the development of effective public policies related to these issues. In Brazil, despite their significantly influence over the society, many government interventions in this segment are undertaken without the adequate assessment of their impacts. This happen because the importance of these tools is not always recognized or because proper tools are not available. The goal of this study was to develop a statistically consistent prognostic tool of road fuel consumption and emissions, which could be used for this purpose. Initially, by surveying the evolutionary trends and known alternatives related to fuels, vehicular technology and transportation systems, it was examined all the aspects that should be considered for this work. Next it was reviewed the main domestic and foreign public models and tools of this sector. Realizing that they could not be used, due to the limits imposed by data availability, it is proposed a new approach to achieve this goal. First it is developed a set of worksheets integrating all data and calculations of a bottom-up model, similar to the ones used in national inventories of toxic emissions. The set was interconnected and configured to facilitate, through an iterative process, the fine tuning of the uncertain estimates, in such way that the model total consumptions as much as possible reproduce the fuel consumptions observed in the country. Next econometric models were developed to estimate total fuel consumptions based on identified relevant economic indicators. And inserting these econometric models in the previous bottom-up model, it was generated a hybrid model that allows inventorying and forecasting of fuel consumptions and vehicular emissions segregated by classes. When compared with the observed fuel consumptions, these models presented statistically robust results. Under some conditions, these results can be converted in CO2 emissions. The same cannot be said with respect to other emissions, partially due to the difficulty to link emission sources to air quality measurements, which does not invalidate the use of the hybrid model to obtain comparative results for these emissions. Finally, the models were submitted to a sensitivity analysis and their applicability was verified for some scenarios.

Les réseaux ad hoc de véhicules accueillent une multitude d’applications intéressantes. Parmi celles-ci, les applications d’info-divertissement visent à améliorer l’expérience des passagers. Ces applications ont des exigences rigides en termes de délai de livraison et de débit. De nombreuses approches ont été proposées pour assurer la qualité du service des dites applications. Elles sont réparties en deux couches : réseau et contrôle d’accès. Toutefois, ces méthodes présentent plusieurs lacunes. Cette thèse a trois volets. Le premier aborde la question du routage dans le milieu urbain. A cet égard, un nouveau protocole, appelé SCRP, a été proposé. Il exploite l’information sur la circulation des véhicules en temps réel pour créer des épines dorsales sur les routes et les connecter aux intersections à l’aide des nœuds de pont. Ces derniers collectent des informations concernant la connectivité et le délai, utilisées pour choisir les chemins de routage ayant un délai de bout-en-bout faible. Le deuxième s’attaque au problème d’affectation des canaux de services afin d’augmenter le débit. A cet effet, un nouveau mécanisme, appelé ASSCH, a été conçu. ASSCH collecte des informations sur les canaux en temps réel et les donne à un modèle stochastique afin de prédire leurs états dans l’avenir. Les canaux les moins encombrés sont sélectionnés pour être utilisés. Le dernier volet vise à proposer un modèle analytique pour examiner la performance du mécanisme EDCA de la norme IEEE 802.11p. Ce modèle tient en compte plusieurs facteurs, dont l’opportunité de transmission, non exploitée dans IEEE 802.11p. / The fact that vehicular ad hoc network accommodates two types of communications, Vehicle-to-Vehicle and Vehicle-to-Infrastructure, has opened the door for a plethora of interesting applications to thrive. Some of these applications, known as infotainment applications, focus on enhancing the passengers' experience. They have rigid requirements in terms of delivery delay and throughput. Numerous approaches have been proposed, at medium access control and routing layers, to enhance the quality of service of such applications. However, existing schemes have several shortcomings. Subsequently, the design of new and efficient approaches is vital for the proper functioning of infotainment applications. This work proposes three schemes. The first is a novel routing protocol, labeled SCRP. It leverages real-time vehicular traffic information to create backbones over road segments and connect them at intersections using bridge nodes. These nodes are responsible for collecting connectivity and delay information, which are used to select routing paths with low end-to-end delay. The second is an altruistic service channel selection scheme, labeled ASSCH. It first collects real-time service channels information and feeds it to a stochastic model that predicts the state of these channels in the near future. The least congested channels are then selected to be used. The third is an analytical model for the performance of the IEEE 802.11p Enhanced Distributed Channel Access mechanism that considers various factors, including the transmission opportunity (TXOP), unexploited by IEEE 802.11p.

Las ciudades altamente congestionadas requieren de una óptima gestión de la movilidad urbana. Muchas de estas ciudades se caracterizan porque su infraestructura vial no va acorde con su demanda vehicular, como es el caso de la ciudad de Lima. Por consiguiente, se debe priorizar el desarrollo de la infraestructura vial, ya que existen por ejemplo intersecciones amplias que presentan gran cantidad de puntos de conflicto. En efecto, reducen la capacidad operacional y el nivel de servicio de la intersección. Esta investigación tiene como objetivo reducir los puntos de conflicto y el tiempo de cruce en intersecciones amplias tipo cruz, en base a la Metodología de Flujo Continuo (CFI, por sus siglas en inglés) con delineadores elevados (bolardos compuestos) y detectores (sensores) a fin de mejorar la capacidad y seguridad operacional de la intersección. El objeto de estudio de la presente investigación se encuentra en una zona comercial ubicada entre la Avenida República de Venezuela con la Avenida Aurelio García y García, Lima. Los movimientos vehiculares de giro a la izquierda son los que afectan el rendimiento de las intersecciones señalizadas. Por tanto, se recolectaron 3219 vehículos en total, de los cuales 561 vehículos giraron a la izquierda demostrando así la alta densidad de este tipo de maniobras. La efectividad del diseño propuesto se valida empleando una microsimulación en el software Vissim. Los resultados obtenidos muestran los puntos de conflictos tipo cruz se redujeron a 8 puntos con el modelo propuesto frente al estado actual, el cual posee 58 puntos de conflictos tipo cruz. Por otro lado, el tiempo promedio de cruce de los vehículos disminuyó en 4.20 y 9.79 segundos en la fase 1, debido al rediseño de la geometría de la intersección. En efecto, la capacidad operacional aumentó en 33.85%, ya que el aforo vehicular con el nuevo ciclo semafórico de 150 segundos es de 431vehículos con respecto a los 322 vehículos que se tenía con el modelo anterior. Finalmente, los riesgos de colisiones posteriores y laterales se reducen mediante el uso de delineadores elevados en un 83%. / Highly congested cities require optimal management of urban mobility. Many of these cities are characterized because their road infrastructure does not match their vehicular demand, as is the case in the city of Lima. Therefore, the development of road infrastructure should be prioritized, since there are, for example, wide intersections that present a large number of conflict points. In effect, they reduce the operational capacity and service level of the intersection. This research aims to reduce conflict points and crossing time at wide cross-type intersections, based on the Continuous Flow Methodology (CFI) with elevated eyeliners (composite bollards) and detectors (sensors) in order to improve the capacity and safety of the intersection. The object of study of the present investigation is in a commercial area located between avenue Republica de Venezuela with avenue Aurelio Garcia y Garcia, Lima. The left-turn vehicular movements are those that affect the performance of the signposted intersections. Therefore, 3219 vehicles were collected in total, of which 561 vehicles turned left demonstrating the high density of this type of maneuver. The effectiveness of the proposed design is validated using microsimulation in the Vissim software. The results obtained show the points of cross-type conflicts were reduced to 8 points with the proposed model compared to the current state, which has 58 points of cross-type conflicts. On the other hand, the average crossing time of the vehicles decreased by 4.20 and 9.79 seconds in phase 1, due to the redesign of the intersection geometry. In fact, the operational capacity increased by 33.85%, since the vehicle capacity with the new traffic light cycle of 150 seconds is 431 vehicles with respect to the 322 vehicles that had the previous model. Finally, the risks of posterior and lateral collisions are reduced by the use of elevated eyeliners by 83%. / Tesis

En la actualidad la congestión vehicular es un problema de muchas ciudades del mundo. Este problema no solo se ha convertido en un sinónimo de estrés y ansiedad para algunas personas, sino que también se ha convertido en un reto para los expertos en Pavimentos. La congestión no solo implica un alto incremento de carga sobre las vías, sino que también implica cambios en las propiedades mecánicas del pavimento. La velocidad vehicular, temperatura y la sobrecarga del pavimento dentro del tráfico son componentes que sin duda se deben ver reflejados en una metodología, ya que estos están alterando la vida útil de las vías.  La presente investigación propone una metodología empírica de determinación del módulo de elasticidad en una carpeta asfáltica. Dentro de esta nueva propuesta se incluyen conceptos que no se han tomado en cuenta: la influencia de la congestión. El experimento de comparación se realizó entre una vía a temperatura de congestión y sin congestión, 35°C y 25 °C respectivamente. Este método fue verificado mediante el uso de fórmulas y ensayos de módulo de elasticidad. / Currently, traffic congestion is a problem in many cities around the world. This problem has not only become synonymous with stress and anxiety for some people, but it has also become a challenge for flooring experts. Congestion not only implies a high increase in load on the roads, but also implies changes in the mechanical properties of the pavement. Vehicle speed, temperature and pavement overload within traffic are components that should undoubtedly be reflected in a methodology, since these are altering the useful life of the roads. This research proposes an empirical methodology for determining the modulus of elasticity in an asphalt mat. This new proposal includes concepts that have not been considered: the influence of congestion. The comparison experiment was carried out between a pathway at a congestion temperature and without congestion, 35 ° C and 25 ° C, respectively. This method was verified by using formulas and modulus of elasticity tests. / Trabajo de investigación

Efforts towards standardization of Vehicular Communication Systems (VCSs) have been conclusive on the use of Vehicular Public-Key Infrastructure (VPKI) for the establishment of trust among network participants. Employing VPKI in Vehicular Communication (VC) guarantees the integrity and authenticity of Cooperative Awareness Messages (CAMs) and Decentralized Environmental Notification Messages (DENMs). It also offers a level of privacy for vehicles as VPKI provides them with a set of non-linkable short-lived certificates, called pseudonyms, which are used to sign outgoing messages by vehicles while they communicate with other vehicles referred to as Vehicle-to-Vehicle (V2V) or Roadside Units (RSUs) referred to as Vehicle-to-Infrastructure (V2I). Each vehicle uses a pseudonym for its lifetime and by switching to a not- previously- used pseudonym, it continues to communicate without risking its privacy. There have been two approaches suggested by the literature on how to provide vehicles with pseudonyms. One is the so-called pre-loading mode, suggesting to pre-load vehicles with all pseudonyms they need, which increases the cost of revocation in case they are compromised. The other one is the on-demand mode, suggesting a real-time offering of pseudonyms by VPKI at vehicles request e.g., on starting each trip. Choosing the on-demand approach imposes a considerable burden of availability and resilience on VPKI services. In this work, we are confronting the problems regarding a large-scale deployment of an on-demand VPKI that is resilient, highly available, and dynamically scalable. In order to achieve that, by leveraging state-of-the-art tools and design paradigms, we have enhanced a VPKI system to ensure that it is capable of meeting enterprise-grade Service Level Agreement (SLA) in terms of availability, and it can also be cost-efficient as services can dynamically scale-out in the presence of high load, or possibly scale-in when facing less demand. That has been made possible by re-architecting and refactoring an existing VPKI into a cloud-native solution deployed as microservices. Towards having a reliable architecture based on distributed microservices, one of the key challenges to deal with is Sybil-based misbehavior. By exploiting Sybil-based attacks in VPKI, malicious vehicles can gain influential advantage in the system, e.g., one can affect the traffic to serve its own will. Therefore, preventing the occurrence of Sybil attacks is paramount. On the other hand, traditional approaches to stop them, often come with a performance penalty as they verify requests against a relational database which is a bottleneck of the operations. We propose a solution to address Sybil-based attacks, utilizing Redis, an in-memory data store, without compromising the system efficiency and performance considerably. Running our VPKI services on Google Cloud Platform (GCP) shows that a large-scale deployment of VPKI as a Service (VPKIaaS) can be done efficiently. Conducting various stress tests against the services indicates that the VPKIaaS is capable of serving real world traffic. We have tested VPKIaaS under synthetically generated normal traffic flow and flash crowd scenarios. It has been shown that VPKIaaS managed to issue 100 pseudonyms per request, submitted by 1000 vehicles where vehicles kept asking for a new set of pseudonyms every 1 to 5 seconds. Each vehicle has been served in less than 77 milliseconds. We also demonstrate that, under a flash crowd situation, with 50000 vehicles, VPKIaaS dynamically scales out, and takes ≈192 milliseconds to serve 100 pseudonyms per request submitted by vehicles. / Ansträngningar för standardisering av Vehicular Communication Systems har varit avgörande för användandet av Vehicular Public-Key Infrastructure (VPKI) för att etablera förtroende mellan nätverksdeltagare. Användande av VPKI i Vehicular Communication (VC) garanterar integritet och autenticitet av meddelanden. Det erbjuder ett lager av säkerhet för fordon då VPKI ger dem en mängd av icke länkbara certifikat, kallade pseudonym, som används medan de kommunicerar med andra fordon, kallat Vehicle-to-Vehicle (V2V) eller Roadside Units (RSUs) kallat Vehicle-to-Infrastructure (V2I). Varje fordon använder ett pseudonym under en begränsad tid och genom att byta till ett icke tidigare använt pseudonym kan det fortsätta kommunicera utan att riskera sin integritet. I litteratur har två metoder föreslagits för hur man ska ladda fordon med pseudonym de behöver. Den ena metoden det så kallade offline-läget, som proponerar att man för-laddar fordonen med alla pseudonym som det behöver vilket ökar kostnaden för revokering i fall de blir komprometterat. Den andra metoden föreslår ett on-demand tillvägagångssätt som erbjuder pseudonym via VPKI på fordonets begäran vid början av varje färd. Valet av på begäran metoden sätter en stor börda på tillgänglighet och motståndskraft av VPKI tjänster. I det här arbetet, möter vi problem med storskaliga driftsättningar av en på begäran VPKI som är motståndskraftig, har hög tillgänglighet och dynamiskt skalbarhet i syfte att uppnå dessa attribut genom att nyttja toppmoderna verktyg och designparadigmer. Vi har förbättrat ett VPKI system för att säkerställa att det är kapabelt att möta SLA:er av företagsklass gällande tillgänglighet och att det även kan vara kostnadseffektivt eftersom tjänster dynamiskt kan skala ut vid högre last eller skala ner vid lägre last. Detta har möjliggjorts genom att arkitekta om en existerande VPKI till en cloud-native lösning driftsatt som mikrotjänster. En av nyckelutmaningarna till att ha en pålitlig arkitektur baserad på distribuerade mikrotjänster är sybil-baserad missuppförande. Genom att utnyttja Sybil baserade attacker på VPKI, kan illvilliga fordon påverka trafik att tjäna dess egna syften. Därför är det av största vikt att förhindra Sybil attacker. Å andra sidan så dras traditionella metoder att stoppa dem med prestandakostnader. Vi föreslår en lösning för att adressera Sybilbaserade attacker genom att nyttja Redis, en in-memory data-store utan att märkbart kompromissa på systemets effektivitet och prestanda. Att köra våra VPKI tjänster på Google Cloud Platform (GCP) och genomföra diverse stresstester mot dessa har visat att storskaliga driftsättningar av VPKI as a Service (VPKIaaS) kan göras effektivt samtidigt som riktigt trafik hanteras. Vi har testat VPKIaaS under syntetisk genererat normalt trafikflöde samt flow och flash mängd scenarier. Det har visat sig att VPKIaaS klarar att utfärda 100 pseudonym per förfråga utsänt av 1000 fordon (där fordonen bad om en ny uppsättning pseudonym varje 1 till 5 sekunder), och varje fordon fått svar inom 77 millisekunder. Vi demonstrerar även att under en flashcrowd situation, där antalet fordon höjs till 50000 med en kläckningsgrad på 100. VPKIaaS dynamiskt skalar ut och tar ≈192 millisekunder att betjäna 100 pseudonymer per förfrågan gjord av fordon.

Esta tesis propone un escenario de solución ante la existente congestión vehicular en el cruce de la avenida Velasco Astete con avenida Primavera situada en el distrito de San Borja límite con Surco. Teniendo en cuenta la demanda y oferta vehicular actual, y las características geométricas que esta intersección y sus calles colindantes poseen. Asimismo, se evaluó el comportamiento del cruce estudiado bajo el recién implementado plan de restricción vehicular “pico y placa”. Dicho plan efectivamente alteraba el flujo vehicular, a pesar de no ser una vía afectada por esta medida, se concluyó que afectaba indirectamente el flujo vehicular. Cabe resaltar que el único propósito de evaluar las condiciones de este plan de restricción vehicular era para recolectar la información de campo, con la mayor veracidad posible y obtener resultados exactos y reales. Una vez recogida toda la información en campo, se analizaron diversos parámetros, para poder obtener valores como el nivel de servicio y las demoras de la intersección en estudio. Habiendo obtenido los resultados de las condiciones actuales de la vía, se evaluaron los posibles planes de mitigación para mejorar las condiciones de servicio del área de estudio. Optando por un rediseño geométrico en colaboración con una semaforización coordinada en el cruce de la avenida de la Floresta y calle Monte Pío con avenida Primavera. Se obtuvo una mejora notoria tanto para el nivel de servicio como para las demoras, en el cruce estudiado, sin tener que sacrificar la fluidez existente en el cruce nuevo aperturado. / This thesis proposes a solution scenario to the existing vehicular congestion at the intersection of Velasco Astete Avenue and Primavera Avenue located in the district of San Borja bordering Surco. Considering the current demand and vehicle offer, and the geometric characteristics that this intersection and its adjoining streets have. Likewise, the behavior of the studied crossing was evaluated under the newly implemented “Pico y Placa” vehicle restriction plan. This plan effectively altered the vehicle flow, despite not being a road affected by this measure, it was concluded that it indirectly affected the vehicle flow. It should be noted that the only purpose of evaluating the conditions of this vehicle restriction plan was to collect the field information, as accurately as possible and obtain accurate and real results. Once all the information was collected in the field, various parameters were analyzed, in order to obtain values ​​such as the level of service and the delays of the intersection under study. Having obtained the results of the current road conditions, the possible mitigation plans were evaluated to improve the service conditions of the study area. Opting for a geometric redesign in collaboration with a coordinated traffic light at the intersection of Avenida de la Floresta and Monte Pío Street with Primavera Avenue. A noticeable improvement was obtained in the studied crossing, not only for level of service but for the delays too, without having to sacrifice the fluidity existing in the newly opened crossing. / Tesis

La presente investigación está orientada a mejorar los niveles de servicio de una rotonda mediante la optimización de su capacidad, a partir de cambios en los elementos de diseño geométrico. En ese sentido se realizó un análisis y evaluación de la rotonda “Higuereta”, la cual se encuentra ubicada en el distrito de Santiago de Surco, en la ciudad de Lima en Perú. Para ello, se realizó una recopilación de bases teóricas a nivel internacional para conocer los fundamentos del diseño geométrico. Es así que, partiendo de un modelo matemático determinístico proveniente de Reino Unido, Kimber 1980, que asocia los parámetros geométricos con el flujo circulante y demuestra que cualquier cambio en un elemento de diseño geométrico es directamente proporcional a la capacidad, podemos concluir que existe mejoras en los niveles de servicio siempre que realicemos variaciones significativas en la geometría. En la presente investigación se evidenció mejoras de un nivel de servicio “F” hasta uno “D”. Por otro lado, se modeló la rotonda a partir de un software de micro simulación, PTV Vissim 9, con el objetivo de poder reflejar mediante simulaciones el comportamiento real del tráfico antes y después de introducir cambios en los parámetros geométricos. Se concluyó que existe mejoras en los niveles de servicio por acceso, de un “D” a un “C. Finalmente, para tener la certeza que el modelo reflejó la realidad se calibró a través de ajustes en los parámetros de comportamiento del conductor y se validó a partir de un indicador estadístico. / This research is aimed at improving the service levels of a roundabout by optimizing its capacity, based on changes in the geometric design elements. In this sense, an analysis and evaluation of the “Higuereta” roundabout was carried out, which is located in the district of Santiago de Surco, in the city of Lima in Peru. For this, a compilation of theoretical bases was carried out at an international level to know the fundamentals of geometric design. Thus, based on a deterministic mathematical model from the United Kingdom, Kimber 1980, which associates geometric parameters with circulating flow and shows that any change in a geometric design element is directly proportional to capacity, we can conclude that there are improvements. in service levels as long as we make significant variations in geometry. In the present investigation, improvements from a service level “F” to a “D” were evidenced. On the other hand, the roundabout was modeled from a micro simulation software, PTV Vissim 9, with the aim of being able to reflect through simulations the real traffic behavior before and after introducing changes in the geometric parameters. It was concluded that there are improvements in service levels per access, from a “D” to a “C. Finally, to be certain that the model reflected reality, it was calibrated through adjustments in the driver's behavior parameters and validated from a statistical indicator. / Tesis

[pt] A perda de percurso (PL) é um parâmetro essencial em modelos de propagação e crucial na determinação da área de cobertura de sistemas móveis. Os métodos de aprendizado de máquina (ML) tornaram-se ferramentas promissoras para a previsão de propagação de rádio. No entanto, ainda existem alguns desafios para sua implantação completa, relacionados à seleção das entradas mais significativas do modelo, à compreensão de suas contribuições para as previsões do modelo e à avaliação adicional da capacidade de generalização para amostras desconhecidas. Esta tese tem como objetivo projetar modelos de PL baseados em ML otimizados para diferentes aplicações das tecnologias 5G e além. Essas aplicações abrangem links de ondas milimétricas (mmWave) para ambientes indoor e outdoor na faixa de frequência de 26,5 a 40 GHz, cobertura de macrocélulas no espectro sub-6 GHz e comunicações veiculares usando campanhas de medições desenvolvidas em CETUC, Rio de Janeiro, Brazil. Vários algoritmos de ML são explorados, como redes neurais artificiais (ANN), regressão de vetor de suporte (SVR), floresta aleatória (RF) e aumento de árvore de gradiente (GTB). Além disso, estendemos dois modelos empíricos para mmWave com previsão de PL melhorada. Propomos uma metodologia para seleção robusta de modelos de ML e uma metodologia para selecionar os preditores mais adequados para as máquinas consideradas com base na melhoria de desempenho e na interpretabilidade do modelo. Além disso, para o canal veículo-veículo (V2V), uma técnica de rede neural convolucional (CNN) também é proposta usando uma abordagem de aprendizado por transferência para lidar com conjuntos de dados pequenos. Os testes de generalização propostos mostram a capacidade dos modelos de ML de aprender o padrão entre as entradas do modelo e a PL, mesmo em ambientes e cenários mais desafiadores de amostras desconhecidas. / [en] Path loss (PL) is an essential parameter in propagation models and critical in determining mobile systems’ coverage area. Machine learning (ML) methods have become promising tools for radio propagation prediction. However, there are still some challenges for its full deployment, concerning to selection of the most significant model s inputs, understanding their contributions to the model s predictions, and a further evaluation of the generalization capacity for unknown samples. This thesis aims to design optimized ML-based PL models for different applications for the 5G and beyond technologies. These applications encompass millimeter wave (mmWave) links for indoor and outdoor environments in the frequency band from 26.5 to 40 GHz, macrocell coverage in the sub-6 GHz spectrum, and vehicular communications using measurements campaign carried out by the Laboratory of Radio-propagation, CETUC, in Rio de Janeiro, Brazil. Several ML algorithms are exploited, such as artificial neural network (ANN), support vector regression (SVR), random forest (RF), and gradient tree boosting (GTB). Furthermore, we have extended two empirical models for mmWave with improved PL prediction. We proposes a methodology for robust ML model selection and a methodology to select the most suitable predictors for the machines considered based on performance improvement and the model’s interpretability. In adittion, for the vehicle-to-vehicle (V2V) channel, a convolutional neural network (CNN) technique is also proposed using a transfer learning approach to deal with small datasets. The generalization tests proposed shows the ability of the ML models to learn the pattern between the model’s inputs and PL, even in more challenging environments and scenarios of unknown samples.