Submitted by (ana.araujo@ufrpe.br) on 2016-08-03T14:07:18Z No. of bitstreams: 1 Luiz Jose Rodrigues dos Santos.pdf: 1081987 bytes, checksum: 435fc2cb438881b9c3905e16b4b41ed0 (MD5) / Made available in DSpace on 2016-08-03T14:07:42Z (GMT). No. of bitstreams: 1 Luiz Jose Rodrigues dos Santos.pdf: 1081987 bytes, checksum: 435fc2cb438881b9c3905e16b4b41ed0 (MD5) Previous issue date: 2008-02-28 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Urban traffic represents a phenomenon of great socioeconomic importance,whose modeling from the point of view of prevision on the basis of initial conditions, still represents a challenge for modern science. Computational methods (computer simulations) represent a powerful tool for modeling and prediction of a number of effects, where systems of coupled differential equations may be used to simulate different phenomena observed in traffic systems. In particular, a quantity of high importance for maintenance and planning of road systems is the vehicular capacity which can be supported without traffic jams, whose description and prevision is still not well understood. In this work, a generalization of an existing microscopic traffic model, the Intelligent Driver Model (IDM), is proposed by implementing a distribution of desired velocities, where it is shown that vehicle capacity of multiple lane roads can be measured in a rather realistic manner, as a function of model parameters,which may be adjusted to real observations. / O tráfego urbano representa um fenômeno de grande importância sócio econômica, cuja modelagem de ponto de vista de previsão a partir de condições iniciais, ainda representa um desafio para a ciência moderna. Métodos computacionais (simulação computacional) representam uma ferramenta poderosa para modelagem e previsão de diversos efeitos, nos quais sistemas de equações diferenciais acopladas podem simular diversos fenômenos observados no sistema de tráfego. Em particular, uma grandeza de alto impacto para o gerenciamento e planejamento de rodovias é a capacidade veicular que elas podem suportar sem que aconteça o efeito de congestionamento, cuja descrição e previsão ainda não estão bem entendida. Neste trabalho, propõe-se uma generalização de um modelo microscópico computacional existente, o Intelligent Driver Model (IDM), aplicando uma distribuição de velocidades desejadas, onde torna-se possível medir de forma bastante realista a capacidade veicular de rodovias com múltiplas faixas, em função de parâmetros de modelo, que podem ser ajustados às observações reais.

Car following (CF) behaviour modelling has made significant progress in both traffic engi-neering and traffic psychology during recent decades. Autonomous vehicles (AVs) have been demonstrated to optimise traffic flow and increase traffic stability. Consequently, sever-al car-following models have been proposed based on various car following criteria, leading to a range of model parameter sets. In traffic engineering, Intelligent Driving Model (IDM) are commonly used as microscopic traffic flow models to simulate a single vehicle's behav-iour on a road. Observational data can be employed to parameter calibrate IDM models, which enhances their practicality for real-world applications. As a result, the calibration of model parameters is crucial in traffic simulation research and typically involves solving an optimization problem. Within the given context, the Nelder-Mead（NM）algorithm, particle swarm optimization (PSO) algorithm and genetic algorithm (GA) are utilized in this study for parameterizing the IDM model, using abundant trajectory data from five different road conditions. The study further examines the effects of various algorithms on the IDM model in different road sections, providing useful insights for traffic simulation and optimization.:Table of Contents CHAPTER 1 INTRODUCTION 1 1.1 BACKGROUND AND MOTIVATION 1 1.2 STRUCTURE OF THE WORK 3 CHAPTER 2 BACKGROUND AND RELATED WORK 4 2.1 CAR-FOLLOWING MODELS 4 2.1.1 General Motors model and Gazis-Herman-Rothery model 5 2.1.2 Optimal velocity model and extended models 6 2.1.3 Safety distance or collision avoidance models 7 2.1.4 Physiology-psychology models 8 2.1.5 Intelligent Driver model 10 2.2 CALIBRATION OF CAR-FOLLOWING MODEL 12 2.2.1 Statistical Methods 13 2.2.2 Optimization Algorithms 14 2.3 TRAJECTORY DATA 21 2.3.1 Requirements of Experimental Data 22 2.3.2 Data Collection Techniques 22 2.3.3 Collected Experimental Data 24 CHAPTER 3 EXPERIMENTS AND RESULTS 28 3.1 CALIBRATION PROCESS 28 3.1.1 Objective Function 29 3.1.2 Errors Analysis 30 3.2 SOFTWARE AND METHODOLOGY 30 3.3 NM RESULTS 30 3.4 PSO RESULTS 37 3.4.1 PSO Calibrator 37 3.4.2 PSO Results 44 3.5 GA RESULTS 51 3.6 OPTIMIZATION PERFORMANCE ANALYSIS 58 CHAPTER 4 CONCLUSION 60 REFERENCES 62

This thesis tackles the problem of predicting the collision risk for vehicles driving in complex traffic scenes for a few seconds into the future. The method is based on previous research using dynamic Bayesian networks to represent the state of the system. Common risk prediction methods are often categorized into three different groups depending on their abstraction level. The most complex of these are interaction-aware models which take driver interactions into account. These models often suffer from high computational complexity which is a key limitation in practical use. The model studied in this work takes interactions between drivers into account by considering driver intentions and the traffic rules in the scene. The state of the traffic scene used in the model contains the physical state of vehicles, the intentions of drivers and the expected behaviour of drivers according to the traffic rules. To allow for real-time risk assessment, an approximate inference of the state given the noisy sensor measurements is done using sequential importance resampling. Two different measures of risk are studied. The first is based on driver intentions not matching the expected maneuver, which in turn could lead to a dangerous situation. The second measure is based on a trajectory prediction step and uses the two measures time to collision (TTC) and time to critical collision probability (TTCCP). The implemented model can be applied in complex traffic scenarios with numerous participants. In this work, we focus on intersection and roundabout scenarios. The model is tested on simulated and real data from these scenarios. %Simulations of these scenarios is used to test the model. In these qualitative tests, the model was able to correctly identify collisions a few seconds before they occur and is also able to avoid false positives by detecting the vehicles that will give way. / Detta arbete behandlar problemet att förutsäga kollisionsrisken för fordon som kör i komplexa trafikscenarier för några sekunder i framtiden. Metoden är baserad på tidigare forskning där dynamiska Bayesianska nätverk används för att representera systemets tillstånd. Vanliga riskprognosmetoder kategoriseras ofta i tre olika grupper beroende på deras abstraktionsnivå. De mest komplexa av dessa är interaktionsmedvetna modeller som tar hänsyn till förarnas interaktioner. Dessa modeller lider ofta av hög beräkningskomplexitet, vilket är en svår begränsning när det kommer till praktisk användning. Modellen som studeras i detta arbete tar hänsyn till interaktioner mellan förare genom att beakta förarnas avsikter och trafikreglerna i scenen. Tillståndet i trafikscenen som används i modellen innehåller fordonets fysiska tillstånd, förarnas avsikter och förarnas förväntade beteende enligt trafikreglerna. För att möjliggöra riskbedömning i realtid görs en approximativ inferens av tillståndet givet den brusiga sensordatan med hjälp av sekventiell vägd simulering. Två olika mått på risk studeras. Det första är baserat på förarnas avsikter, närmare bestämt att ta reda på om de inte överensstämmer med den förväntade manövern, vilket då skulle kunna leda till en farlig situation. Det andra riskmåttet är baserat på ett prediktionssteg som använder sig av time to collision (TTC) och time to critical collision probability (TTCCP). Den implementerade modellen kan tillämpas i komplexa trafikscenarier med många fordon. I detta arbete fokuserar vi på scerarier i korsningar och rondeller. Modellen testas på simulerad och verklig data från dessa scenarier. I dessa kvalitativa tester kunde modellen korrekt identifiera kollisioner några få sekunder innan de inträffade. Den kunde också undvika falsklarm genom att lista ut vilka fordon som kommer att lämna företräde.