Modeling and software tools for freeway operational planning

This dissertation grew out of the author's participation in the Tools for Operational Planning (TOPl) project. TOPl project started at Berkeley in April 2006. Its purpose is to provide tools for quantitative assessment of operational strategies designed to improve traffic conditions on congested freeways and surrounding arterials. The elements of such strategies are: (1) demand management, which focuses on reducing excessive demand; (2) incident management, which targets resources to alleviate accident hot spots; (3) traveler information , which potentially reduces traveler buffer time; and (4) traffic control, which implements aggressive ramp metering at locations where significant reductions in congestion are likely to occur.; TOPl models traffic flows on road networks in realistic conditions relying on freeway data provided by the California freeway Performance Measurement System (PeMS). PeMS offers an accounting framework for tracking freeway performance, and a suite of diagnostic tools that pinpoint the weaknesses in freeway operations. For arterial data, TOP] uses various alternative sources, such as regional planning agencies---MTC (Bay Area) and SANDAG (San Diego) for GIS description of road networks, and census information and demographics surveys for origin-destination travel patterns and demand.; Macroscopic traffic models represent traffic as a compressible fluid in terms of flow, density and speed, as opposed to microscopic models which seek to reproduce the behavior of individual vehicle as its driver responds to its environment by changing its speed and lane. TOPl uses the macroscopic approach, as it is based on sound theory, is easy to implement in software, and the implementations are fast to run, allowing the user to simulate many different traffic situations in relatively short time. Our model of choice is the cell transmission model (CTM)---a special case of Godunov discretization of the Lighthill-Whitham-Richards first order model, with triangular fundamental diagram. While simple, CTM adequately describes traffic flow on freeways, and the simulation results match well the measurement data provided by PeMS. Exploring the CTM model from the point of view of nonlinear dynamical systems we describe the structure of its equilibrium points and the behavior of its trajectories under different demand patterns. Then we discuss the implications of our findings for ramp metering.; CTMSIM is the interactive MATLAB based freeway traffic simulator of the CTM model. It allows plugging in user-defined ramp flow and ramp queue controllers, its output results can be directly compared with PeMS data, and it can operate in both graphical (interactive) mode and command line (batch) mode. Simple and robust, CTMSIM has proved to be a handy tool for transportation researchers who can use it for evaluating ramp metering algorithms and for estimating the impact of different response times in the incident management. CTMSIM can only deal with a single freeway. It does not support arterials, freeway networks, or even HOV lanes.; The Aurora object-oriented framework overcomes the limitations of the CTMSIM. Its basic objects, nodes and links, allow the user to construct heterogeneous road networks. Various event classes make it possible to generate simulation scenarios. The monitor objects can keep track of the state at selected nodes and links, coordinate control actions at nodes, or generate events at nodes or links when the monitored states reach certain thresholds. Monitors and events enable the modeling of the impact of traveler information as well as incident management and the coordination of signal control on arterials with the ramp metering at freeway entries. The analysis module of Aurora, which is still under development, will address the issue of demand management: the goal is to solve the user equilibrium dynamic traffic assignment problem and evaluate various toll mechanisms.; There are other benefits of Aurora.; It is...