Research Of Urban Traffic Network Signal Iterative Learning Control Based On Macroscopic Fundamental Diagram

Traffic congestion has become one of the most challenging problems of city management in modern China due to the high economic growth,the fast u rbanization process and the increases of vehicle population.Traffic congestion may not only lead to more traffic accidents and decrease the traffic service levels,but may cause the environmental pollution and energy waste as well.Different traffic management measures are implemented in order to alleviate traffic congestion problems in large cities,and among many others,traffic signal control is the most promising and convenient way to achieve the goal.With the rapid development of new intelligent control technology,such as computer technology,automatic control,artificial intelligence and so on,it is a hot spot in the research of intelligent traffic signal control system to alleviate traffic congestion thro ugh intelligent traffic control theories and techniques.Most of the existing signal control methods which rely on hierarchical control structure are model-based ones.However,due to the uncertainties in the real traffic network a nd even the occurrence of un-modeled traffic dynamics,the effectiveness of the model-based control may be weakened.It is evident that there exist certain repeatable patterns in the urban traffic network,e.g.,the morning commute causes the peak hours in weekdays,and normally,plenty of historical traffic data is available.So we propose a partially data-driven signal control strategy for urban traffic network which can be formulated into the hierarchical control structure.Especially,we gather the traffic data to achieve the Macroscopic Fundamental Diagram(MFD),which characterizes the general statistical properties of the traffic network and provides the preferable control objective,and we use iterative learning control(ILC)method to achieve the signal timing plan for the network.The main contents of the thesis are summarized as follows:1.The general behavior of each subregion in the traffic network is described and a coordinated control scheme based on MFD is proposed.The criterions based on Normalized Cuts for dividing the subregions are given,and certain subregion partitioning algorithm is provided.For each subregion,field data and simulation data are collected to demonstrate the existence and the variation of MFD.Moreover,by using the curve of MFD,the optional vehicle occupancy of road network is calculated and it serves as the control objective for the signal control design in the lower layer.2.The lower layer proposes a signal control strategy of sub area network based on ILC.The real network of each subregion is modeled,and by using ILC method,the appropriate signal timing plan can be achieved which satisfies the control objective.The convergence of the proposed ILC-based algorithm is proved.3.The effectiveness of the algorithm is verified through Matlab and Vissim simulations.The comparison analysis is also carried out with respect to the Webster fixed timing scheme.Finally,we analyze and summarize the research contents of the article.