Connected Transportation Systems: Next Generation Traffic Simulation and Data Collection Tools and Techniques =Ba?l? Ula?t?rma Sistemleri: Yeni Jenerasyon Trafik Simülasyon ve Veri Toplama Araçlar? ve Teknikler

A connected environment is where a user, an application, or a node of the system is continuously connected to another node or nodes. Considering every element of this environment is also a data provider, connected environments promote mesh networking which enables hundreds and even thousands of devices that communicate with each other without the need of central communication hubs. The main principle of connected environments is to achieve a higher goal which could not be achieved by any of the environment's individual elements without exchanging their information and interacting with the individual elements of the system. While a connected environment will change the way we imagine and design our mutual relationship to natural and built systems, it is still in its early days. In this context, this Dissertation studies several real-life and simulated connected environments. It evaluates the quality of data generated in each system by using big data analytics and attempts to understand their impacts on reliability, safety, and mobility of transportation systems.;Transportation is undergoing a revolution driven by this idea of connectivity briefly mentioned above. Earlier Chapters in this dissertation mainly explore the opportunities to simulate connectivity within transportation networks and to build applications to improve traffic mobility as well as safety. To achieve this goal, microscopic traffic simulation tools coupled with customized plugins for simulating communication scenarios are used. The primary performance measure is selected to be travel time to examine the reliability of state estimations using the data generated by these connected vehicles. Machine learning algorithms that were implemented using hierarchical clustering methodology improved the accuracy of the estimation of performance measures for even lower market penetrations.;In near future, with the help of decentralized connected systems, real-time coordination and information-sharing, drivers will have the opportunity of putting themselves in control of their own destiny. Therefore, it is imperative to understand the speed of transferring information from one vehicle to all the others, especially in the presence of accidents where the information dissemination time becomes a crucial issue. This dissertation proposes an analytical framework based on a disease-spread model to estimate the time it takes to transfer the critical information to the target network. This proposed method does not require the development of a detailed transportation network using traditional traffic simulation tools. Moreover it obviates the need for conducting time-consuming simulation runs that can be prohibitive under certain circumstances especially when the simulated networks are very large and complex. The results show that the developed model can predict the information transfer time reasonably well for higher market penetration levels that are more than 20% and the approach is practical for dense urban scenarios with high traffic densities.;In attempt to move for a simulated environment to the real-world implementation, alternative open data collection procedures for transportation analysis are also introduced. The primary objective here is to acquire and use data for segments in a transportation network where physical sensor infrastructure is limited. The results show that utilizing open data sources can deliver useful information for regular and breakdown traffic conditions. For post-evaluation of traffic incidents, the possibility to examine the impact of incidents on roadways, clearance times, and crowd-sourced incident information is also shown.;Following chapters focus on the integration and the usage of the various ubiquitous Internet of Things (IoT) devices in a connected transportation environment. Sensors developed as part of this work can detect devices with wireless capabilities within a predefined area. In connected environments, IoT sensors will provide a rich amount of real-time data to facilitate the communication among agents of the transportation system. Some of these data include crowd densities, wait-times, and origin-destination flows are acquired and then processed. Initial tests revealed promising results in terms of employing this data for system performance evaluation. Since the data are collected at a major transit hub, passenger arrival behavior is also analyzed to understand and infer the intra-and-inter-daily variation of the passenger arrival intensity. The results illustrate that the arrival intensity at bus terminals is indeed a doubly stochastic process with time-varying intensity.;Latest developments in connected environments have led an exponential growth in data production. While some challenges of data analytics are addressed by big data approaches, structure, and analytics; one must carefully evaluate the integration, implementation, and interface related issues in these emerging connected environments with the goal of improving people's daily commutes and thus overall lives. The findings in this dissertation demonstrate the usefulness and reliability of connected systems for improving transportation operations, traffic mobility and safety. (Abstract shortened by ProQuest.).