Avec la demande croissante d'espace urbain, de plus en plus de parkings à plusieurs étages sont nécessaires. Bien que ces parkings contribuent à une utilisation plus efficace de l'espace urbain, ils introduisent également un nouveau problème. Les rapports suggèrent environ 70 millions d'heures de recherche d'emplacements de stationnement chaque année, soit une perte équivalente de 700 millions d'euros pour la seule France. En outre, les utilisations des parkings vont au-delà de leurs objectifs initiaux. Des fonctionnalités exigeantes telles que le chargeur électrique, la réservation en ligne de places de stationnement, le guidage dynamique ou le paiement mobile, etc. transforment un parking en un environnement intelligent et compétitif. Une solution à ce problème consiste à développer un système de navigation autonome pour les véhicules intelligents en situation de parking. La thèse identifiera une de ces sous-tâches, à savoir la localisation dans des environnements non autorisés par GPS. Cette thèse présentera une nouvelle méthode pour résoudre le problème indiqué tout en maintenant le système en respectant quatre critères: disponibilité, évolutivité, universalité et précision. Il y a deux étapes principales: (1) une solution permettant de reproduire le comportement du GPS pour un environnement refusé par GPS, et (2) un cadre permettant la fusion de systèmes de type GPS avec d'autres méthodes de localisation pour obtenir une précision de localisation élevée. Tout d'abord, un système de localisation Wi-Fi Fingerprinting est utilisé. Une approche utilisant un réseau de neurones d'ensemble sur une base de données d'empreintes hybrides Wi-Fi est proposée dans cette thèse. Des expériences menées sur une durée d'un an montrent que ce système est capable de localiser des véhicules présentant une erreur moyenne de 2,25 m dans le repère global (WGS84). Deuxièmement, une solution de localisation complète doit être une fusion de plusieurs techniques. Cela permet aux niveaux de localisation global et local de fonctionner ensemble. Parallèlement, la redondance dans le système améliore la précision et la fiabilité. Dans cette thèse, un cadre de fusion flexible pour plusieurs capteurs de localisation est proposé. Ce cadre de fusion traitera non seulement de l'environnement refusé par le GPS, mais pourrait également être utilisé dans l'environnement assisté par GPS et assurer une transition en douceur entre les deux zones. Pour accomplir cette tâche exigeante, un filtre à particules modèle de mélange gaussien est développé. Alors que le modèle de mouvement de ce filtre à particules intègre des données provenant de l'IMU (unité de mesure inertielle) ou du laser-SLAM, le modèle de correction est un modèle de mélange gaussien de plusieurs observations obtenues à partir du système de localisation d'empreintes digitales Wi-Fi. Avec deux véhicules intelligents (une Cybercar et une Citroen C1), 64 expériences ont été réalisées pour valider le cadre. Une erreur de localisation moyenne de 0,5 m est obtenue dans un cadre de coordonnées global. Comparez avec d'autres solutions avec une erreur de localisation moyenne de 0,2 m dans les cadres de coordonnées locales; Cette solution proposée présente également des avantages en termes d'évolutivité, de disponibilité et d'universalité. / With the increasing demand for urban space, more and more multistory carparks are needed. Although these carparks help to utilize urban space more efficient, they also introduce a new problem. Reports suggest approximately 70 million hours of parking slot searching each year, equivalently 700 million euros loss for France alone. In addition, carparks uses are exceeding their original purposes. Demanding features such as electric charger, online booking of parking spaces, dynamic guidance or mobile payment etc. turn a carpark into a competitive smart environment. One solution to this problem is to develop an autonomous navigation system for intelligent vehicles in the carpark situation. The thesis will identify one of these sub-tasks namely localization in GPS-denied environments. This thesis will present a novel method to solve the indicated problem while keeping the system follows four criteria: availability, scalability, universality and accuracy. There are two main steps: (1) a solution to replicate the GPS behaviour for the GPS-denied environment, and (2) a framework that allows the fusion of GPS-like systems with other localization methods to achieve a high localization accuracy. First, a Wi-Fi Fingerprinting localization system is employed. An approach using an ensemble neural network on a hybrid Wi-Fi fingerprinting database is proposed in this thesis. Experiments in a year-long duration show that this system is capable of localizing vehicles with 2.25m of mean error in the global coordinate frame (WGS84). Second, a complete localization solution must be a fusion of multiple techniques. This allows global as well as local levels of localization to function together. At the same time, having redundancy in the system boosts accuracy and reliability. In this thesis, a flexible fusion framework for multiple localization sensors is proposed. This fusion framework will not only deal with the GPS-denied environment but could be potentially used in the GPS-aided environment and provide a smooth transition between the two areas. To accomplish this demanding task, a Gaussian Mixture Model Particle Filter is developed. While the motion model of this particle filter incorporates data from the IMU (Inertial Measurement Unit) or laser-SLAM, the correction model is a Gaussian mixture model of multiple observations obtained from the Wi-Fi fingerprinting localization system. With two intelligent vehicles (a Cybercar and a Citroen C1 car), 64 experiments were carried out to validate the framework. A mean localization error of 0.5m is achieved in a global coordinate frame. Compare to other solutions with 0.2m of mean localization error in local coordinate frames; this proposed solution has advantages in terms of scalability, availability and universality as well.

Master of Science

Les transports terrestres occupent une place majeure dans notre société, notamment en ville où les ralentissements aux heures de pointe peuvent avoir un impact notable sur l'organisation des activités, l'économie ou encore l'écologie. Les infrastructures routières sont généralement coordonnées par un centre de contrôle, responsable du maintien des équipements, de leurs réglages initiaux et de la gestion des incidents (matériels ou humains). Les nouvelles technologies de l'information et de la communication ont permis, en l'espace de quelques années, de mettre en œuvre des systèmes de transport intelligents. À l'aide de multiples points de mesures répartis sur le territoire, un opérateur peut dénombrer les usagers et en déduire la charge du réseau. Toutefois, centraliser les informations présente de nombreuses limites. Cette thèse vise à étudier l'emploi de systèmes distribués afin de mettre en œuvre des systèmes de transport intelligents grâce à un réseau de capteurs sans fil. Couplés à une unité de détection (p. ex., un magnétomètre), les capteurs communicants peuvent réagir au passage d'un véhicule en étant déployés, par exemple, sur les voies. Ils sont également capables de coopérer et de s'affranchir d'une entité centrale, rendant tout ou partie d'une zone urbaine totalement indépendante. D'autre part, ces réseaux peuvent fonctionner de manière autonome et tolèrent mieux les pannes, car aucun élément n'est indispensable au fonctionnement global du système. Enfin, les éléments de ces réseaux sont petits, peu coûteux, et communiquent en sans fil, ce qui leur permet d'être déployés et redéployés rapidement et de manière dense. / Road traffic has a significant effect on metropolitan activities, especially during peak hours when it impacts on areas such as the economy and the environment. Road infrastructure is typically coordinated from a control centre that is responsible for maintaining not only its equipment but also their initial settings and incident management (both material and human). During the past few years, new technologies in the fields of information and communication have led to the introduction of intelligent transportation systems. Using multiple measurement points distributed across a country, an operator can count road users and calculate the network load. However, the centralization of information has a number of drawbacks. The aim of this thesis is to study the use of distributed systems in order to implement intelligent transportation systems via a wireless sensor network. Coupled to a detection unit (e.g., a magnetometer), the interconnected sensors can respond to the passage of a vehicle when deployed, for example, along the road. They can also work together without recourse to a central entity - rendering all or part of an urban area totally independent. Furthermore, these networks can operate autonomously and are less susceptible to breadown, because the overall running of the system is not affected by the failure of individual components. Finally, components are small and cheap, and they operate wirelessly, which means they can be deployed and redeployed both rapidly and densely.

With the rapid improvement of the Advanced Driver Assistant System (ADAS), autonomous driving has become one of the most common hot topics in recent years. While driving, many technologies related to autonomous driving choose to use the sensors installed on the vehicle to collect the information of road status and the environment outside. This aims to warn the driver to perceive the potential danger in the fastest time, which has become the focus of autonomous driving in recent years. Although autonomous driving brings plenty of conveniences to people, the safety of it is still facing difficulties. During driving, even the experienced driver can not guarantee focus on the status of the road all the time. Thus, lane departure warning system (LDWS) becomes developed. The purpose of LDWS is to determine whether the vehicle is in the safe driving area. If the vehicle is out of this area, LDWS will detect it and alert the driver by the sensors, such as sound and vibration, in order to make the driver back to the safe driving area. This thesis proposes a novel lightweight LDWS model LEHA, which divides the entire LDWS into three stages: image preprocessing, lane detection, and lane departure recognition. Different from the deep learning methods of LDWS, our LDWS model LEHA can achieve high accuracy and efficiency by relying only on simple hardware. The image preprocessing stage aims to process the original road image to remove the noise which is irrelevant to the detection result. In this stage, we apply a novel algorithm of grayscale preprocessing to convert the road image to a grayscale image, which removes the color of it. Then, we design a binarization method to greatly extract the lane lines from the background. A newly-designed image smoothing is added to this stage to reduce most of the noise, which improves the accuracy of the following lane detection stage. After obtaining the processed image, the lane detection stage is applied to detect and mark the lane lines. We use region of interest (ROI) to remove the irrelevant parts of the road image to reduce the detection time. After that, we introduce the Canny edge detection method, which aims to extract the edges of the lane lines. The last step of LDWS in the lane detection stage is a novel Hough transform method, the purpose of it is to detect the position of the lane and mark it. Finally, the lane departure recognition stage is used to calculate the deviation distance between the vehicle and the centerline of the lane to determine whether the warning needs to turn on. In the last part of this paper, we present the experiment results which show the comparison results of different lane conditions. We do the statistic of the proposed LDWS accuracy in terms of detection and departure. The detection rate of our proposed LDWS is 98.2% and the departure rate of it is 99.1%. The average processing time of our proposed LDWS is 20.01 x 10⁻³s per image.

Safety in transportation systems is of paramount concern to society; many improvements have been made in recent decades and yet thousands of fatalities still occur annually. Work zones in particular are areas with increased safety risks in transit networks. Advances in electronics have now allowed engineers to merge powerful computing and communication technologies with modern automotive and vehicular technology, known as connected vehicle. Connected vehicle will allow vehicles to exchange data wirelessly with each other and infrastructure to improve safety, mobility and sustainability. This thesis presents a paper that focuses on evaluating the impact of connected vehicle on work zone traffic safety. A dynamic route guidance system based on decaying average-travel-time and shortest path routing was developed and tested in a microscopic traffic simulation environment to avoid routes with work zones. To account for the unpredictable behaviour and psychology of driver’s response to information, three behaviour models, in the form of multinomial distributions, are proposed and studied in this research. The surrogate safety measure improved Time to Collision was used to gauge network safety at various market penetrations of connected vehicles. Results show that higher market penetrations of connected vehicles decrease network safety due to increased average travel distance, while the safest conditions, 5%-10% reduction in critical Time to Collision events, were observed at market penetrations of 20%-40% connected vehicle, with network safety strongly influenced by behaviour model. / Thesis / Master of Applied Science (MASc)

Over the past few years, the Utah Department of Transportation (UDOT) has developed a system called the Signal Performance Metrics System (SPMS) to evaluate the performance of signalized intersections. This system currently provides data summaries for several performance measures including: 1) Purdue Coordination Diagram, 2) Speed, 3) Approach Volume, 4) Purdue Phase Termination Charts, 5) Split Monitor, 6) Turning Movement Volume Counts, 7) Arrivals on Red, and 8) Approach Delay. There is a need to know the accuracy of the data that are being collected by the Wavetronix SmartSensor Matrix and displayed in the SPMS. The TAC members determined that the following factors would affect the accuracy of radar-based traffic sensors the most: sensor position, number of approach lanes, and volume level. The speed limit factor was added to the study after most of the data collection was completed. The purpose of this research was to collect data at various intersections to determine the accuracy of the data collected by the Wavetronix SmartSensor Matrix.A Mixed Model Analysis of Variance (ANOVA) was employed to analyze the effects that each factor had on the accuracy of the traffic volume count. A total of 14 tests were performed to examine the effects of the factors on traffic volume count accuracy. The sensor position factor was not found to be a statistically significant factor affecting the accuracy of traffic volume counts. The effect of speed limit on traffic volume count accuracy was determined to be inconclusive due to the lack of samples to be tested. The remaining two factors, volume level and number of approach lanes, were found to have a statistically significant effect on the accuracy of traffic volume counts. Based on these two factors, a matrix was created to meet the needs of UDOT to present accuracy values on the SPMS website. This matrix includes the mean, 95 percent confidence interval of the mean, standard deviation, number of samples, and the minimum number of samples needed.

This research evaluates Intelligent Transportation System (ITS) implementation strategies to improve the safety of a freeway once a potential of a crash is detected. Among these strategies are Variable Speed Limit (VSL) and ramp metering. VSL are ITS devices that are commonly used to calm traffic in an attempt to relieve congestion and enhance throughput. With proper use, VSL can be more cost effective than adding more lanes. In addition to maximizing the capacity of a roadway, a different aspect of VSL can be realized by the potential of improving traffic safety. Through the use of multiple microscopic traffic simulations, best practices can be determined, and a final recommendation can be made. Ramp metering is a method to control the amount of traffic flow entering from on-ramps to achieve a better efficiency of the freeway. It can also have a potential benefit in improving the safety of the freeway. This thesis pursues the goal of a best-case implementation of VSL. Two loading scenarios, a fully loaded case (90% of ramp maximums) and an off-peak loading case (60% of ramp maximums), at multiple stations with multiple implementation methods are strategically attempted until a best-case implementation is found. The final recommendation for the off-peak loading is a 15 mph speed reduction for 2 miles upstream and a 15 mph increase in speed for the 2 miles downstream of the detector that shows a high crash potential. The speed change is to be implemented in 5 mph increments every 10 minutes. The recommended case is found to reduce relative crash potential from .065 to -.292, as measured by a high-speed crash prediction algorithm (Abdel-Aty et al. 2005). A possibility of crash migration to downstream and upstream locations was observed, however, the safety and efficiency benefits far outweigh the crash migration potential. No final recommendation is made for the use of VSL in the fully loaded case (low-speed case); however, ramp metering indicated a promising potential for safety improvement.

Commuters in urban areas suffer from traffic congestion on a daily basis. The increasing number of vehicles and vehicle miles traveled (VMT) are exacerbating this congested roadway problem for society. Although literature contains numerous studies that strive to propose solutions to this congestion problem, the problem is still prevalent today. Traffic congestion problem affects society’s quality of life socially, economically, and environmentally. In order to alleviate the unsustainable impacts of the congested roadway problem, Intelligent Transportation Systems (ITS) has been utilized to improve sustainable transportation systems in the world. The purpose of this thesis is to analyze the sustainable impacts and performance of the utilization of ITS in the United States. This thesis advances the body of knowledge of sustainability impacts of ITS related congestion relief through a triple bottom line (TBL) evaluation in the United States. TBL impacts analyze from a holistic perspective, rather than considering only the direct economic benefits. A critical approach to this research was to include both the direct and the indirect environmental and socio-economic impacts associated with the chain of supply paths of traffic congestion relief. To accomplish this aim, net benefits of ITS implementations are analyzed in 101 cities in the United States. In addition to the state level results, seven metropolitan cities in Florida are investigated in detail among these 101 cities. For instance, the results of this study indicated that Florida saved 1.38 E+05 tons of greenhouse gas emissions (tons of carbon dioxide equivalent), $420 million of annual delay reduction costs, and $17.2 million of net fuel-based costs. Furthermore, to quantify the relative impact and sustainability performance of different ITS technologies, several ITS solutions are analyzed in terms of total costs (initial and operation & maintenance costs) and benefits (value of time, emissions, and safety). To account for the uncertainty in benefit and cost ii analyses, a fuzzy-data envelopment analysis (DEA) methodology is utilized instead of the traditional DEA approach for sustainability performance analysis. The results using the fuzzy-DEA approach indicate that some of the ITS investments are not efficient compared to other investments where as all of them are highly effective investments in terms of the cost/benefit ratios approach. The TBL results of this study provide more comprehensive picture of socio-economic benefits which include the negative and indirect indicators and environmental benefits for ITS related congestion relief. In addition, sustainability performance comparisons and TBL analysis of ITS investments contained encouraging results to support decision makers to pursue ITS projects in the future.

In this thesis, multiple on-road vehicle tracking problem is explored, with greater consideration of road constraints and interactions between vehicles. A comprehensive method for tracking multiple on-road vehicles is proposed by making use of domain knowledge of on-road vehicle motion. Starting with raw measurements provided by sensors, bias correction methods for sensors commonly used in vehicle tracking are briefly introduced and a fast but effective bias correction method for airborne video sensor is proposed. In the proposed method, by assuming errors in sensor parameter measurements are close to zero, the bias is separately addressed in converted measurements of target position by a linear term of errors in sensor parameter measurements. Based on this model, the bias is efficiently estimated by addressing it while tracking or using measurements of targets that are observed by multiple airborne video sensors simultaneously. The proposed method is compared with other airborne video bias correction methods through simulations. The numerical results demonstrate the effectiveness of the proposed method for correcting bias as well as its high computational efficiency. Then, a novel tracking algorithm that utilizes domain knowledge of on-road vehicle motion, i.e., road-map information and interactions among vehicles, by integrating a car-following model into a road coordinate system, is proposed for tracking multiple vehicles on single-lane roads. This algorithm is extended for tracking multiple vehicles on multi-lane roads: The road coordinate system is extended to two-dimension to express lanes on roads and a lane-changing model is integrated for modeling lane-changing behavior of vehicles. Since the longitudinal and lateral motions are mutually dependent, the longitudinal and lateral states of vehicles are estimated sequentially in a recursive manner. Two estimation strategies are proposed: a) The unscented Kalman filter combined with the multiple hypothesis tracking framework to estimate longitudinal and lateral states of vehicles, respectively. b) A unified particle filter framework with a specifically designed computationally-efficient joint sampling method to estimate longitudinal and lateral states of vehicles jointly. Both of two estimation methods can handle unknown parameters in motion models. A posterior Cramer-Rao lower bound is derived for quantifying achievable estimation accuracy in both single-lane and multi-lane cases, respectively. Numerical results show that the proposed algorithms achieve better track accuracy and consistency than conventional multi-vehicle tracking algorithms, which assumes that vehicles move independently of one another. / Thesis / Doctor of Philosophy (PhD)

In the realm of autonomous vehicles, ensuring safety through advanced anomaly detection is crucial. This thesis integrates Reservoir Computing with temporal-aware data analysis to enhance driver behavior assessment and traffic flow prediction. Our approach combines Reservoir Computing with autoencoder-based feature extraction to analyze driving metrics from vehicle sensors, capturing complex temporal patterns efficiently. Additionally, we extend our analysis to forecast traffic flow dynamics within road networks using the same framework. We evaluate our model using the PEMS-BAY and METRA-LA datasets, encompassing diverse traffic scenarios, along with a GPS dataset of 10,000 taxis, providing real-world driving dynamics. Through a support vector machine (SVM) algorithm, we categorize drivers based on their performance, offering insights for tailored anomaly detection strategies. This research advances anomaly detection for autonomous vehicles, promoting safer driving experiences and the evolution of vehicle safety technologies. By integrating Reservoir Computing with temporal-aware data analysis, this thesis contributes to both driver behavior assessment and traffic flow prediction, addressing critical aspects of autonomous vehicle systems. / Master of Science / Our cities are constantly growing, and traffic congestion is a major challenge. This project explores how innovative technology can help us predict traffic patterns and develop smarter management strategies. Inspired by the rigorous safety systems being developed for self-driving cars, we'll delve into the world of machine learning. By combining advanced techniques for identifying unusual traffic patterns with tools that analyze data over time, we'll gain a deeper understanding of traffic flow and driver behavior. We'll utilize data collected by car sensors, such as speed and turning patterns, to not only predict traffic jams but also see how drivers react in different situations. However, our project has a broader scope than just traffic flow. We aim to leverage this framework to understand driver behavior in general, with a particular focus on its implications for self-driving vehicles. Through meticulous data analysis and sophisticated algorithms, we can categorize drivers based on their performance. This valuable information can be used to develop improved methods for detecting risky situations, ultimately leading to safer roads and smoother traffic flow for everyone. To ensure the effectiveness of our approach, we'll rigorously test it using real-world data from GPS data from taxi fleets and nationally recognized traffic datasets. By harnessing the power of machine learning and tools that can adapt to changing data patterns, this project has the potential to revolutionize traffic management in cities. This paves the way for a future with safer roads, less congestion, and a more positive experience for everyone who lives in and travels through our bustling urban centers.