Date of Award

8-2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Mechanical Engineering

Committee Member

Dr. Ardalan Vahidi, Committee Chair

Committee Member

Dr. Mohammed Daqaq

Committee Member

Dr. John R. Wagner

Committee Member

Dr. Mashrur Chowdhury

Abstract

Vehicular traffic flow is essential, yet complicated to analyze. It describes the interplay among vehicles and with the infrastructure. A better understanding of traf-fic would benefit both individuals and the whole society in terms of improving safety, energy efficiency, and reducing environmental impacts. A large body of research ex-ists on estimation and control of vehicular traffic in which, however, vehicles were assumed not to be able to share information due to the limits of technology. With the development of wireless communication and various sensor devices, Connected Vehicles(CV) are emerging which are able to detect, access, and share information with each other and with the infrastructure in real time. Connected Vehicle Technology (CVT) has been attracting more and more attentions from different fields. The goal of this dissertation is to develop approaches to estimate and control vehicular traffic as well as individual vehicles relying on CVT. On one hand, CVT sig-nificantly enriches the data from individuals and the traffic, which contributes to the accuracy of traffic estimation algorithms. On the other hand, CVT enables commu-nication and information sharing between vehicles and infrastructure, and therefore allows vehicles to achieve better control and/or coordination among themselves and with smart infrastructure. The first part of this dissertation focused on estimation of traffic on freeways and city streets. We use data available from on road sensors and also from probe One of the most important traffic performance measures is travel time. How-ever it is affected by various factors, and freeways and arterials have different travel time characteristics. In this dissertation we first propose a stochastic model-based approach to freeway travel-time prediction. The approach uses the Link-Node Cell Transmission Model (LN-CTM) to model traffic and provides a probability distribu-tion for travel time. The probability distribution is generated using a Monte Carlo simulation and an Online Expectation Maximization clustering algorithm. Results show that the approach is able to generate a reasonable multimodal distribution for travel-time. For arterials, this dissertation presents methods for estimating statistics of travel time by utilizing sparse vehicular probe data. A public data feed from transit buses in the City of San Francisco is used. We divide each link into shorter segments, and propose iterative methods for allocating travel time statistics to each segment. Inspired by K-mean and Expectation Maximization (EM) algorithms, we iteratively update the mean and variance of travel time for each segment based on historical probe data until convergence. Based on segment travel time statistics, we then pro-pose a method to estimate the maximum likelihood trajectory (MLT) of a probe vehicle in between two data updates on arterial roads. The results are compared to high frequency ground truth data in multiple scenarios, which demonstrate the effectiveness of the proposed approach. The second part of this dissertation emphasize on control approaches enabled by vehicular connectivity. Estimation and prediction of surrounding vehicle behaviors and upcoming traffic makes it possible to improve driving performance. We first propose a Speed Advisory System for arterial roads, which utilizes upcoming traffic

Share

COinS