Joint Queue Estimation And Max Pressure Control For Signalized Urban Networks With Connected Vehicles

Urban traffic congestion is a global problem.Advanced traffic signal control is an important measure to alleviate traffic congestion,while traffic state estimation sets the foundation for adaptive signal control.Max Pressure(MP)is a decentralized adaptive signal control strategy,which highly depends on the real-time information of vehicle queues in urban networks.This paper studies queue estimation and adaptive traffic signal control based on the MP algorithm using connected vehicle dataConnected vehicle data is obtained with the GPS technology,and available GPS data for civil usage is not accurate enough to locate vehicles by lane.So,it is impossible to distinguish between straight&right-turn(S&RT)queues and left-turn(LT)queue at a signalized junction using current connected vehicle data.On the other hand,the S&RT and LT queues at a signalized intersection are typically managed by different signal phases.The distinction between the two queues is crucial for adaptive traffic signal control,e.g.that using the MP strategy.Thus,a dilemma arises.Two heuristic MP strategies with connected vehicle data are proposed for real-time link queue estimation and traffic signal control.For comparison,an ideal(future)situation is also considered,where connected vehicles can be directly located by lane,so that the two types of queues can be readily distinguished to fascilitate lane-wise queue estimation and the MP traffic signal controlIn this thesis,we have developed three adaptive signal control plans along with a fixed-time plan.The queue estimation algorithm and control plans were evaluated on an AIMSUN-based microscopic simulation platform with respect to a 3*3 urban network under two scenarios involving various connected vehicle proportionsThe results show that the performance of queue estimation and MP control under the ideal situation is satisfactory,while there is a performance gap between the practical situation and ideal situation.As the proportion of connected vehicles increases,the performance of the queue estimation and MP signal control algorithm improves.