This dissertation explores issues in the control and optimization of Intelligent Transportation Systems, with particular emphasis on the design of dynamic routing and traffic signal control strategies in local and freeway traffic networks.; To formulate the control and optimization problems in traffic networks, representations of static and dynamic information of traffic networks are provided. We use a set of topology matrices to represent the traffic static information and give two dynamic models, a point-queuing model and a hydrodynamic model, to represent the characteristics of traffic in local and freeway networks.; Based on the point-queuing model, we first analyze the characteristics of dynamic routing and traffic signal control in local traffic networks. First, a decentralized multi-destination dynamic routing strategy is proposed, in which the values of the cost-to-go of every link to different destinations in a traffic network are calculated in real time to direct traffic. Then, a decentralized hybrid intersection controller with five states is designed to address the different combinations of traffic situations at an intersection, the downstream and upstream intersections in a network and the problem that queues may overrun an upstream intersection.; Based on the hydrodynamic model, we analyze the dynamic routing and signal control in freeway traffic networks. Using the sliding mode design concept, we propose a dynamic two-freeway routing method and an on-ramp controller. This on-ramp controller can iteratively search for the optimal operating point under the given conditions and at the same time meter inflow to achieve this desired state.
展开▼
