PhD Student: Yunjie Zhao
Publication Year: 2014
Advisor: Adel Sadek
Abstract: To address transportation externalities (e.g., congestion, accidents, pollution, depletion of energy resources), the concept of Intelligent Transportation Systems, and its latest evolution known as the Connected Vehicles (CV) initiative, has been recently proposed. CVs utilize cyber technologies for sensing, communications and networking, to enable unprecedented levels of connectivity among vehicles, infrastructure and travelers. Besides connectivity, future transportation systems (hereby referred to as Cyber Transportation Systems or CTS) will exhibit high levels of automation that will eventually result in partial and fully autonomous vehicle control. While CTS provide tremendous opportunities for all transportation stakeholders to rethink and redefine how the transportation system works, there is currently an urgent need for next generation modeling platforms to address the design, testing and evaluation of CTS applications. Motivated by such a need, Motivated by such a need, this dissertation was dedicated to the development of three simulation frameworks, including a large-scale agent-based modeling (ABM) traffic simulator, an integrated traffic-emission simulator, and an integrated traffic-driving-network simulator (ITDNS).
Specifically, the large-scale ABM traffic study focuses on modeling a middle-size metropolitan region network from an agent-based perspective. Each individual travel agent has attributes such as its trip origin, destination, travel mode, route choice, driving pattern and so forth. Built upon that, the second phase of that study aimed at capturing riving behavior under inclement weather conditions (specifically snow) and reproducing such an impact in the simulation world. A freeway incident scenario was then considered in order to evaluate the performance of the CTS solution (i.e., the provision of real-time travel and dynamic route guidance information). The evaluations were also escalated to the compounded effect of a freeway incident and inclement weather.
Secondly, the integrated traffic and emission model addresses the critical need to conduct project-level emissions analysis. To that end, the dissertation evaluates the performance of different approaches to integrating the emission model MOVES, recently developed by the Environmental Protection Agency (EPA) and traffic simulators. Generally, the second-by-second vehicle trajectory output from the traffic simulator could, in principal, be used to define the link drive schedule required to run the project-level MOVES analysis. However, the challenge is that the MOVES model required defining a representative vehicle trajectory for each link, because tracking the emissions for individual vehicles was computationally intractable. The accuracy of two aggregation and one sampling methods are evaluated for both freeway links as well as arterial links, and for both Cellular Automata based traffic micro-simulators and for car-following models.
The dissertation also carries out the first attempt to integrating traffic, driving, and communications network simulator, resulting in a unique simulation platform, ITDNS, supporting the design and evaluation of novel CTS and CV solutions. The unique feature which distinguishes ITDNS from other similar models is its ability to address the human factor issue in the design and testing process of CV applications. The study first presents the three major components individually, i.e. the traffic, driving and network simulator and discusses the challenges of integrating these three different types of simulators and how they were addressed.
As the demonstration of the simulator capability and advantage of having "human-in-the-loop", an eco-signal case study is conducted. The concept works as follows: the signal timing information is broadcasted to the approaching connected vehicle, so that the intelligent unit in the vehicle could plan the speed to mitigate the excessive acceleration and hard-braking. Other than the intuitive safety benefits, the fuel consumption and emission savings is also observed.