Public Deposited

The term Connected Vehicle (CV) is broadly used to identify any ‘smart vehicle’ with wireless connectivity to the roadside infrastructure and other vehicles. CVs that can be driven autonomously are called connected autonomous vehicles (CAVs). With real-time communication and data transmission capability, CAVs have the potential to improve the transportation system’s traffic flow, reliability, and safety. However, due to lack of data on CAVs their potential has not been fully realized. Consequently, this dissertation explores the potential of CAV technology in formulating 1) facility type-customized active management and control strategies, 2) assessment techniques, and 3) analytical methods with the purpose of enhancing existing transportation network’s operational capabilities. To explore the potential of CAVs, the dissertation is divided into three components. In the first component, the dissertation focuses on freeways and highways. At this level, study first puts forward a computationally efficient framework to model CAVs using existing simulation tools and relevant data. Using this framework, traffic flow conditions and travel time reliability were studied. Results of this framework show that CAVs can improve traffic flow conditions while increasing the reliability of the system under different demand levels and operational conditions including inclement weather condition. In the second component, the dissertation focuses on the arterial roads in a transportation network. In this component, microscopic traffic flow models were utilized to emulate CAVs on the arterial road and formulate two advanced traffic signal control strategies. Two advanced traffic signal control logics are developed, via V2I and V2V communication. These signals provide synchronized traffic flow on major corridors while keeping the logic decentralized. CAVs compute their travel time delay accumulated on a route. This accumulated delay forms basis of decentralized but coordinated traffic signal strategies. Numerical experiments show that the decentralized but coordinated traffic signal strategies outperform state of the art practices. In the third component, the analysis is conducted at the user-defined path level. A user-defined path can consist of both arterial and freeways. Thus, the study of user-defined path opens an interesting avenue to analyze the combination of arterial and freeways and how they interact with each other. The analysis in this component formulates an innovative network partitioning concept based on the average path-level fundamental relationships among the traffic stream variables. Correspondingly, time-of-day control and management strategies can be tailored to suit specific paths’ operational characteristics. Furthermore, to adequately compute link travel time correlations and accurately determine path-based travel time distributions an analytical model was designed. Analytical form of path travel time variance was devised to correctly capture the spatiotemporal covariance of link travel times. Travel time distributions along the paths defined by users were estimated through solving a convoluting integral of correlated link travel times. Numerical experiments show that the model accurately estimates the travel time distribution along paths. The developed simulation techniques, control strategies, and assessment methods should be used to enhance model calibration and validation; enrich system’s performance, performance evaluation; and provide a sound basis for making routing decisions taking quantifiable risk estimates into account.

Last modified
  • 11/29/2018
Date created
Resource type
Rights statement