Perimeter Control Method Of Urban Traffic Sub-region Based On Iterative Learning Control

The rapid growth of car ownership has caused an imbalance in the relationship between urban traffic supply and demand,and the resulting traffic congestion has gradually expanded,which has gradually developed into a regional congestion problem.It has been proven that the new traffic infrastructure or the implementation of policy control measures can not truly solve the regional traffic congestion.Therefore,seeking more reasonable and effective traffic signal control strategies has become the development trend of the study of regional traffic congestion.The paper is titled “Research on Iterative Learning Perimeter Control Method of Urban Traffic Sub-regions”,and mainly studies the application of iterative learning control method in urban traffic perimeter control when considering random disturbances and the nonlinear characteristics of traffic flow,and gives a rigorous proof of convergence.The main work of the thesis is as follows:1.For the perimeter control study of large-scale heterogeneous urban road networks,it is necessary to divide them into different control sub-regions according to the traffic flow characteristics.Firstly,the traffic flow data in the study road network are collected and the correlation values between adjacent signal intersections are calculated using the Whitson model and the traffic flow data;then the correlation degree values were clustered based on the fast global K-means spectral clustering algorithm,and the intersections belonging to the same cluster were divided into the same sub-region,and the final division results were obtained by adjusting the boundaries of the sub-regions formed by the clustered intersections according to the constraints;finally,it is verified by simulation experiments that the sub-regions obtained after the division all have macroscopic fundamental diagram characteristics to prove the effectiveness of the sub-region division method.2.Aiming at the problem that most of the existing perimeter control strategies do not consider the impact of various disturbances and uncertain factors in the actual traffic system on the control performance,an iterative learning perimeter control method for traffic sub-regions considering the disturbance is proposed.Firstly,the vehicle balance equation with a disturbance term is established,and the macroscopic fundamental diagram of the controlled sub-region under different degrees of disturbance is plotted according to the VISSIM simulation results,and the corresponding optimal cumulative number of vehicles is calculated as the desired output;secondly,an open-closed-loop PD-type iterative learning control law is constructed for the disturbances,and the convergence of the iterative learning algorithm when the input quantity is limited is analyzed in detail;finally,the simulation test confirmed that the approach can effectively suppress the impact of disturbances on the road network performance.3.Considering that the actual urban traffic flow is strongly nonlinear,this paper is based on the existing research on nonlinear traffic flow,and further constructs a macroscopic traffic flow model with nonlinear terms based on the macroscopic traffic flow inflow and outflow relations of the control sub-region,proposes an iterative learning perimeter control method for the traffic sub-region considering the nonlinear characteristics of the traffic flow,and validates the convergence of the proposed method based on the Lipschitz condition and the definition of partial derivatives;finally,the validity of iterative learning control for complex nonlinear traffic flow control is confirmed by comparative simulation experiments.