Traffic Congestion Control Strategies In Urban Road Networks

Transportation is the main driving force of the development in modern metropolitan cities.Due to the continuous development of urbanization,the rapid growth of traffic demand and the increasing number of vehicles,the contradiction between urban traffic pressure and limited road resources has become more and more serious.As the core of the contradiction,traffic congestion is considered as one of the main problems that degrade urban development.Moreover,traffic congestion not only results in the decline of various economic and social functions,but also leads to the continuous deterioration of environment and further becomes an “urban ailment” hindering development process.With the increasing complexity of China's urban transportation networks,aiming at severe traffic congestion problems,suitable traffic congestion control strategies should be investigated to improve the capability of traffic congestion control,relieve traffic problems,support sustainable city development,and achieve the vision of safe,efficient and green intelligent transportation systems.Accordingly,the following contents are included in this dissertation:1.Aiming at recurrent traffic congestion,with consideration of road congestion degree and congestion propagation effects,this dissertation proposes a bottleneck identification method for urban road networks.Firstly,definitions of traffic bottlenecks and congestion correlation between congestion on different road segments are proposed based on congestion propagation.Then,according to the spatio-temporal relationships among congestion correlations,congestion propagation graphs are constructed to model congestion propagation in urban road networks.After that,the maximum spanning tree and Markov analysis are combined to quantify the impact of congestion propagation of all road segments and identify traffic bottlenecks in road networks.Traffic relief at the identified bottlenecks using the proposed technique can bring more effective network-wide improvement and the proposed method can provide effective suggestions for congestion mitigation and road capacity improvement in the road networks.2.Considering the impacts from both road infrastructure and signal control on traffic congestion,a root cause identification method is developed in this dissertation.Specifically,this dissertation first differentiates congestion effects between road segments and intersections according to the spatial relationships of congestion on road segments and intersections,and proposes the definitions of congestion on road segments and intersections respectively.Then,by analyzing the correlations between congestion on road segments and intersections,a causal congestion tree is constructed to model the congestion propagation pattern in urban road networks.Based on a combined use of graphical models and fuzzy logic,congestion costs of all road segments and intersections in the whole road network are calculated.After that,a Markov model is established to identify the most critical congestion root cause among multiple correlated road segments and intersections.Furthermore,a gradient boosting decision tree based method is presented to predict root cause of congestion according to traffic flows,signal control strategies and road topology in traffic networks.The proposed methods for root cause identification and prediction can attribute the causes of congestion to road infrastructure or signal control,which can provide a specific control strategy for congestion mitigation in urban road networks.3.Aiming at non-recurrent traffic congestion,the dissertation investigates an incident management policy considering both signal control and route choice.Firstly,this dissertation introduces a route choice method according to the proportional switch adjustment process and extends the model with consideration of bottleneck delays.Then,using queuing theory,this dissertation analyzes the route travel costs under incident effects and gives the equilibrium existence condition after the occurrence of an incident.In addition,considering the interactions between route choice and signal control when an incident had been identified,a traffic signal control policy is proposed and the condition for equilibrium existence is given with the consideration of dynamic signal control and route choice simultaneously.The proposed policy can be utilized to improve the efficiency of real-time incident management in urban traffic networks,which is beneficial to improve the recovery rate and stability of road systems under different incident circumstances,and alleviate the negative impacts caused by nonrecurrent traffic congestion.