THESIS
2020
xiv, 163 pages : illustrations (some color) ; 30 cm
Abstract
This thesis investigates adaptive coordinated traffic signal control in the context of demand uncertainty. To address the uncertain nature of traffic demand, we first formulate the adaptive coordinated signal control problem into a two-stage stochastic program. In the first stage, the base signal timing plan, including cycle time, green splits and initial offset, is determined as a long-term plan. In the second stage, upon realizations of traffic demand, adaptive control strategies are applied to respond to occasional overflows. We encapsulate the cell transmission model (CTM) into the two-stage stochastic program to jointly and explicitly address three fundamental traffic flow characteristics, i.e., dynamic, spatial and stochastic, for adaptive signal control. The concept of phase clea...[
Read more ]
This thesis investigates adaptive coordinated traffic signal control in the context of demand uncertainty. To address the uncertain nature of traffic demand, we first formulate the adaptive coordinated signal control problem into a two-stage stochastic program. In the first stage, the base signal timing plan, including cycle time, green splits and initial offset, is determined as a long-term plan. In the second stage, upon realizations of traffic demand, adaptive control strategies are applied to respond to occasional overflows. We encapsulate the cell transmission model (CTM) into the two-stage stochastic program to jointly and explicitly address three fundamental traffic flow characteristics, i.e., dynamic, spatial and stochastic, for adaptive signal control. The concept of phase clearance reliability (PCR) is incorporated into the signal control problem. We propose a PCR-based framework to decouple the intertwined two-stage problem into sub-problems and solve them separately, which greatly enhances the solution efficiency. To enable the implementation of the ‘two-stage’ signal control framework in real-time traffic conditions, we propose a cycle-based rolling horizon optimization scheme. Keeping the base timing plan (BTP) fixed, we optimize offsets and green extensions with the updated traffic arrival and queuing information for each horizon. We then develop a software-in-the-loop system based on VISSIM to evaluate the proposed adaptive coordinated signal control plans. The two-stage stochastic programming framework is also extended to the network-wide signal operations while accounting for travelers’ route choice behavior. In the first stage, the base signal plan with a control buffer against variability is introduced to control the equilibrium flows and the resulting steady-state performance. Upon realizations of the random demand, adaptive signal settings are then determined in the second stage to address the occasional demand overflows so as to avoid transient congestion and spillback. A reliability-based gradient projection algorithm is developed to solve the network-wide signal design problem. Numerical examples are provided to demonstrate the properties of the proposed methods.
Post a Comment