Esse trabalho consiste em um estudo sobre a aplicabilidade da medida de centralidade de intermediação (betweenness centrality) combinada com demandas de tráfego em redes viárias com o objetivo de identificar os principais pontos dessas redes. Como principais pontos refere-se aqui aos que aparecem com maior frequência entre os caminhos utilizados pelos motoristas que se deslocam pela rede viária. Trata-se de um estudo exploratório, que se inicia com a aplicação da centralidade de intermediação sobre redes simples, estendendo-se até simulações sobre redes baseadas em mapas reais. Nesse trabalho é analisado o comportamento da medida de centralidade sobre a topologia da rede - i.e. tanto sem considerar uma demanda, como considerando demandas de diversas magnitudes. Para ilustrar a proposta, os resultados são confrontados com valores de centralidade de intermediação calculados sobre as taxas de ocupação das vias extraídas de simulação microscópica. Ao final, foram apresentadas evidências de que o método proposto consegue explicar os fluxos de tráfego com melhor desempenho do que a centralidade de intermediação original. No entanto, o método mostrou-se muito sensível à função de custo utilizada na atribuição da demanda de tráfego ao grafo da rede. Os melhores resultados demonstrados pela abordagem proposta foram obtidos em experimentos sobre redes não regulares e com demandas de tráfego não uniformes. No caso de redes regulares com demanda uniforme, o melhor desempenho foi obtido pelo cálculo da centralidade sem considerar a demanda, mas atribuindo-se o custo unitário às arestas do grafo representativo da rede. / This work consists of a study of applicability of betweenness centrality combined with traffic demands in road networks with the objective of identifying their central points. By central points we refer to those which appear with high frequency among the paths used by drivers that move along the road network. It is an exploratory study, which begins with the application of the betweenness centrality on simple networks, extending to simulations on networks based on real maps. In this study we have analyzed the behavior of the metric over the network topology - i.e. without considering demand, as well as experiments considering demands with several magnitudes. To illustrate the proposed method, the results are compared with betweenness centrality values calculated over roadways occupation rates extracted from microscopic simulation. At the end, evidence that the proposed method can explain traffic flows with better performance than the original betweenness centrality were presented. However, the proposed method was shown to be very sensitive to the cost function used in the allocation of the graph network traffic demand. The best results demonstrated by the proposed approach were obtained in experiments on nonregular networks and non-uniform traffic demands. In the case of regular networks with uniform demand, the best performance was obtained by calculating the betweenness centrality without considering the demand, but assigning the unitary cost to the edges of the network graph.

Sistemas de transporte são sistemas complexos compostos de diferentes entidades que interagem entre si. A otimização do uso da infraestrutura de transporte existente, que é cada vez mais necessária dado o crescente aumento da demanda por mobilidade, passa pela simulação de novas tecnologias que podem vir a ser utilizadas no futuro, como a comunicação inter-veicular (IVC) e a tarifação viária adaptativa. Esta dissertação apresenta uma abordagem baseada em agentes para simulação de comunicação inter-veicular e tarifação viária adaptativa em sistemas de transporte. Motoristas são modelados como agentes minimizadores de custo, composto pelo tempo de viagem e pelas despesas com tarifas viárias. Os motoristas podem usar IVC para expandir seu conhecimento do estado da rede viária. Entre os motoristas que usam IVC, podem existir agentes maliciosos, que buscam afastar os demais de suas rotas, através da divulgação de informações falsas. Os agentes maliciosos podem ainda agir de maneira coordenada, de modo a divulgarem informações falsas sobre as rotas de todos os agentes do grupo. Pelo lado da infraestrutura, gerentes viários percebem o fluxo de veículos nas vias da rede viária e definem as tarifas a serem aplicadas através de um esquema de aprendizado por reforço. Nos experimentos realizados, empregamos um modelo microscópico de simulação de tráfego, o que permite observar o comportamento individual de cada entidade do sistema de transporte sob estudo. O cenário onde as simulações são executadas é uma rede viária com as principais vias arteriais da cidade de Porto Alegre, Brasil. Resultados experimentais indicam que um pequeno grupo coordenado de agentes maliciosos em cenários de IVC é capaz de causar prejuízos significativos aos demais motoristas. Embora na média o grupo não consiga reduzir seu tempo de viagem, alguns agentes maliciosos são beneficiados pela coordenação do grupo. Com relação à tarifação viária, os resultados experimentais indicam que o esquema de aprendizado por reforço não possui a mesma eficácia de um esquema de tarifação fixa quando se trata da maximização de fluxo de veículos na rede viária. Ambos os esquemas de tarifação são superados por um método de otimização de tráfego que assume conhecimento completo do estado da rede viária pelos motoristas. No aspecto individual, sob tarifação via aprendizado por reforço, os custos de deslocamento dos motoristas são superiores em comparação aos custos sob tarifação fixa. O modelo baseado em agentes apresentado nesta dissertação representa uma contribuição em direção à proposição de uma metodologia para integrar modelos comportamentais de usuários de sistemas de transporte que reagem aos padrões de tráfego e medidas de controle desses padrões, com foco em métodos descentralizados e distribuídos. / Transportation systems are complex systems composed of different interacting entities. The optimization of the existing transportation infrastructure usage, which becomes increasingly necessary given the increasing demand for mobility, requires simulation of new technologies that might be used in the future, such as inter-vehicular communication (IVC) and adaptive road pricing. This dissertation presents an agent-based approach for simulation of inter-vehicular communication and adaptive road pricing in transportation systems. Drivers are modeled as cost-minimizer agents, where the cost is composed by travel time and expenditure. Drivers can use IVC to expand their knowledge of the road network state. Among the IVC users, there might be malicious agents, which try to divert other drivers from their routes by spreading false information. The malicious agents can act in a coordinated way, by spreading false information about the routes of all the agents in the group. In the infrastructure side, link managers perceive the vehicular flow in the roads and define the prices to be applied by means of a reinforcement learning scheme. In the experiments, we employ a microscopic traffic simulation model, which allows us to observe the individual behavior of each entity in the studied transportation system. The scenario where the simulations are run is a road network with the main arterial roads of the city of Porto Alegre, Brazil. Experimental results indicate that a small group of coordinated malicious agents in IVC scenarios is able to cause significant losses to the other drivers. Although in average the group does not succeed in reducing their travel times, some agents are benefited by the coordination of the group. Regarding road pricing, experimental results indicate that the reinforcement learning scheme does not achieve the same effectiveness of a fixed pricing approach regarding the maximization of vehicular flow in the road network. Both pricing schemes are outperformed by an optimization method that assumes full knowledge of the road network state by the drivers. In the individual aspect, under pricing via reinforcement learning, drivers’ costs are higher compared to their costs under fixed pricing. The agent-based model presented in this dissertation is a contribution towards a methodology to integrate behavioral models of human travelers reacting to traffic patterns and control measures of these traffic patterns, focusing on distributed and decentralized methods.

A real-time Vehicle Make and Model Recognition (VMMR) system is a significant component of security applications in Intelligent Transportation Systems (ITS). A highly accurate and real-time VMMR system significantly reduces the overhead cost of resources otherwise required. In this thesis, we present a VMMR system that provides very high classification rates and is robust to challenges like low illumination, occlusions, partial and non-frontal views. These challenges are encountered in realistic environments and high security areas like parking lots and public spaces (e.g., malls, stadiums, and airports). The VMMR problem is a multi-class classification problem with a peculiar set of issues and challenges like multiplicity, inter- and intra-make ambiguity among various vehicles makes and models, which need to be solved in an efficient and reliable manner to achieve a highly robust VMMR system. To reliably overcome the ambiguity challenges, a global features representation approach based on the Bag-of-Features paradigm is proposed. We extract key features from different make-model classes in an optimized dictionary, through two different dictionary building strategies. We represent different samples from each class with respect to the learned dictionary. We also present two classification schemes based on multi-class Support Vector Machines (SVMs): (1) Single multi-class SVM and (2) Attribute Bagging-based Ensemble of multi-class SVMs. These classification schemes allow simultaneous learning of the differences between global representations of different classes and the similarities between different shapes or generations within a same make-model class, to further overcome the multiplicity challenges for real-time application. Extensive experiments conducted using our approaches yield superior results for images that were occluded, under low illumination, partial camera views, or even non-frontal views, available in a recently published real-world VMMR dataset. The approaches presented herewith provide a highly accurate VMMR system for real-time applications in realistic environments.

Providing transportation system operators and travelers with accurate travel time information allows them to make more informed decisions, yielding benefits for individual travelers and for the entire transportation system. Most existing advanced traveler information systems (ATIS) and advanced traffic management systems (ATMS) use instantaneous travel time values estimated based on the current measurements, assuming that traffic conditions remain constant in the near future. For more effective applications, it has been proposed that ATIS and ATMS should use travel times predicted for short-term future conditions rather than instantaneous travel times measured or estimated for current conditions. This dissertation research investigates short-term freeway travel time prediction using Dynamic Neural Networks (DNN) based on traffic detector data collected by radar traffic detectors installed along a freeway corridor. DNN comprises a class of neural networks that are particularly suitable for predicting variables like travel time, but has not been adequately investigated for this purpose. Before this investigation, it was necessary to identifying methods for data imputation to account for missing data usually encountered when collecting data using traffic detectors. It was also necessary to identify a method to estimate the travel time on the freeway corridor based on data collected using point traffic detectors. A new travel time estimation method referred to as the Piecewise Constant Acceleration Based (PCAB) method was developed and compared with other methods reported in the literatures. The results show that one of the simple travel time estimation methods (the average speed method) can work as well as the PCAB method, and both of them out-perform other methods. This study also compared the travel time prediction performance of three different DNN topologies with different memory setups. The results show that one DNN topology (the time-delay neural networks) out-performs the other two DNN topologies for the investigated prediction problem. This topology also performs slightly better than the simple multilayer perceptron (MLP) neural network topology that has been used in a number of previous studies for travel time prediction.

The accurate and reliable estimation of travel time based on point detector data is needed to support Intelligent Transportation System (ITS) applications. It has been found that the quality of travel time estimation is a function of the method used in the estimation and varies for different traffic conditions. In this study, two hybrid on-line travel time estimation models, and their corresponding off-line methods, were developed to achieve better estimation performance under various traffic conditions, including recurrent congestion and incidents. The first model combines the Mid-Point method, which is a speed-based method, with a traffic flow-based method. The second model integrates two speed-based methods: the Mid-Point method and the Minimum Speed method. In both models, the switch between travel time estimation methods is based on the congestion level and queue status automatically identified by clustering analysis. During incident conditions with rapidly changing queue lengths, shock wave analysis-based refinements are applied for on-line estimation to capture the fast queue propagation and recovery. Travel time estimates obtained from existing speed-based methods, traffic flow-based methods, and the models developed were tested using both simulation and real-world data. The results indicate that all tested methods performed at an acceptable level during periods of low congestion. However, their performances vary with an increase in congestion. Comparisons with other estimation methods also show that the developed hybrid models perform well in all cases. Further comparisons between the on-line and off-line travel time estimation methods reveal that off-line methods perform significantly better only during fast-changing congested conditions, such as during incidents. The impacts of major influential factors on the performance of travel time estimation, including data preprocessing procedures, detector errors, detector spacing, frequency of travel time updates to traveler information devices, travel time link length, and posted travel time range, were investigated in this study. The results show that these factors have more significant impacts on the estimation accuracy and reliability under congested conditions than during uncongested conditions. For the incident conditions, the estimation quality improves with the use of a short rolling period for data smoothing, more accurate detector data, and frequent travel time updates.

La presente tesis consiste en la investigación y desarrollo de la implementación de las tecnologías ITS para mejorar la transitabilidad en la intersección de las avenidas Primavera y Velasco Astete. Esta implementación está orientada a la disminución de la congestión vehicular en la intersección a través de una mejor gestión del tráfico sin la necesidad de implementar infraestructuras viales basadas en obras civiles, como ampliación de carriles, creación de viaductos, pasos a desnivel, entre otros, las cuales requieren una mayor inversión de capital. Es por esta razón que el objeto de esta tesis está en la propuesta de la mejora de la situación actual del tráfico en el escenario en estudio con la implementación de las tecnologías ITS mediante su simulación en el software Aimsun v8.2. Para ello, se estudiaron los niveles de servicio según los conceptos establecidos por la metodología HCM 2010. La implementación de los ITS en el escenario estudiado permitió una reducción promedio de 150 segundos en el tiempo de demora, los factores de la cola media y densidad vehicular también presentaron mejoras. Asimismo, el nivel de servicio, luego del análisis de resultados realizado, presentó una mejora significativa al pasar de una clasificación F a una clasificación C en la intersección principal estudiada. Finalmente, del análisis de costos realizado, a través de la asociación a los factores de valor social del tiempo, VHMD y tiempo de viaje, se obtuvo que la implementación de los ITS en la intersección representaría un beneficio de S/. 1,227.77 / km.hora.pasajeros promedio. / This thesis consists of the research and development of the implementation of ITS technologies to improve the passability at the intersection of Primavera and Velasco Astete avenues. This implementation is aimed at reducing vehicular congestion at the intersection through better traffic management without the need to implement road infrastructure based on civil works, such as extension of lanes, creation of viaducts, overpasses, among others, which require a greater capital investment. It is for this reason that the purpose of this thesis is in the proposal to improve the current traffic situation on the stage under study with the implementation of ITS technologies through its simulation in the Aimsun v8.2 software. For this, the service levels were studied according to the concepts established by the HCM 2010 methodology. The implementation of ITS in the scenario studied allowed an average reduction of 150 seconds in the delay time, the factors of the average tail and vehicle density also showed improvements. Likewise, the level of service (LOS), after the analysis of the results, presented a significant improvement when passing from an F classification to a C classification at the main intersection studied. Finally, from the cost analysis carried out, through the association with the social value factors of time, VHMD and travel time, it was obtained that the implementation of STIs at the intersection would represent a benefit of S /. 1,227.77 / km. Hour. Average passengers. / Tesis

abstract: The autonomous vehicle technology has come a long way, but currently, there are no companies that are able to offer fully autonomous ride in any conditions, on any road without any human supervision. These systems should be extensively trained and validated to guarantee safe human transportation. Any small errors in the system functionality may lead to fatal accidents and may endanger human lives. Deep learning methods are widely used for environment perception and prediction of hazardous situations. These techniques require huge amount of training data with both normal and abnormal samples to enable the vehicle to avoid a dangerous situation. The goal of this thesis is to generate simulations from real-world tricky collision scenarios for training and testing autonomous vehicles. Dashcam crash videos from the internet can now be utilized to extract valuable collision data and recreate the crash scenarios in a simulator. The problem of extracting 3D vehicle trajectories from videos recorded by an unknown monocular camera source is solved using a modular approach. The framework is divided into two stages: (a) extracting meaningful adversarial trajectories from short crash videos, and (b) developing methods to automatically process and simulate the vehicle trajectories on a vehicle simulator. / Dissertation/Thesis / Video Demonstration / Masters Thesis Computer Science 2019

Collaborative dispatching allows several dispatchers to view the routing solution as a dynamic model where changes to the vehicle routes can be made in real-time. In this paper we discuss implications of collaborative dispatching on real-time decision support tools for motor carriers. We present a collaborative dispatching system which uses real-time information obtained from a telematics system. Messages sent from the vehicles are automatically analysed and actual data, such as exact arrival and departure times, as well as discrepancies between actual and planned data are identified. The collaborative dispatching system not only allows several dispatchers to concurrently modify the schedule, but also a dynamic optimisation method. The optimisation method is capable of taking into account that input data may change at any time and that dispatchers can concurrently modify the schedule and may add or relax certain constraints relevant to the optimisation model.

Real-life vehicle routing problems encounter a number of complexities that are not considered by the classical models found in the vehicle routing literature. In this paper we consider a dynamic real-life vehicle routing problem which is a combined load acceptance and generalised vehicle routing problem incorporating a diversity of practical complexities. Among those are time window restrictions, a heterogeneous vehicle fleet with different travel times, travel costs and capacity, multi-dimensional capacity constraints, order/vehicle compatibility constraints, orders with multiple pickup, delivery and service locations, different start and end locations for vehicles, route restrictions associated to orders and vehicles, and drivers’ working hours. We propose iterative improvement approaches based on Large Neighborhood Search. Our algorithms are characterised by very fast response times and thus, can be used within dynamic routing systems where input data can change at any time.

While the traditional approach to ease traffic congestion has focused on building infrastructure, the recent emergence of Connected and Automated Vehicles (CAVs) and urban mobility services (e.g., Autonomous Mobility-on-Demand (AMoD) systems) has opened a new set of alternatives for reducing travel times. This thesis seeks to exploit these advances to improve the operation and efficiency of Intelligent Transportation Systems using a network optimization perspective. It proposes novel methods to evaluate the prospective benefits of adopting socially optimal routing schemes, intermodal mobility, and contraflow lane reversals in transportation networks. This dissertation makes methodological and empirical contributions to the transportation domain. From a methodological standpoint, it devises a fast solver for the Traffic Assignment Problem with Side Constraints which supports arbitrary linear constraints on the flows. Instead of using standard column-generation methods, it introduces affine approximations of the travel latency function to reformulate the problem as a quadratic (or linear) programming problem. This framework is applied to two problems related to urban planning and mobility policy: social routing with rebalancing in intermodal mobility systems and planning lane reversals in transportation networks. Moreover, it proposes a novel method to jointly estimate the Origin-Destination demand and travel latency functions of the Traffic Assignment Problem. Finally, it develops a model to jointly optimize the pricing, rebalancing and fleet sizing decisions of a Mobility-on-Demand service. Empirically, it validates all the methods by testing them with real transportation topologies and real traffic data from Eastern Massachusetts and New York City showing the achievable benefits obtained when compared to benchmarks.