Time-of-day Signal Timing Optimization At Isolated Intersection Using License Plate Recognition Data

Time-of-Day signal timing optimization at isolated intersection is an important means to alleviate traffic congestion,improve traffic capacity and reduce traffic delays.Considering that the Time-of-Day signal timing optimization mainly depend on manual regulation and model optimization at domestic intersection,and the accuracy is not high and the optimization effect is not obvious.Using license plate recognition data,this paper proposes a Time-of-Day method and signal timing optimization method that meets the current demand of signal optimization of domestic intersection.In the Time-of-Day method at Isolated Intersection,in view of the problem that the existing algorithm resulting in that time is not continuous and length of the period is too short.The paper proposes a Time-of-Day method based on time series segmentation algorithm.Using license plate recognition data,the time series sample data is extracted by time and space gathering,and the linear regression cost function of the sample data is constructed based on the characteristics of the time series sample data.The dynamic programming search algorithm is constructed with the minimum cost function value as the optimization target.Comparing result of the time series segmentation algorithm with result of K-means clustering algorithm,the results show that the time series segmentation algorithm can solve the problems of time discontinuity and length of time period.Research on timing Optimization at intersection,the index modeling and estimation methods that algorithm minimizes the indicators such as delay and queue length have high complexity.Considering that single ring structure is adopted in traffic signal control,A signal timing method based on stage-phase structure is proposed.The lane-phase mapping relations and stage-phase distribution table are constructed according to the intersection geometry and timing scheme.By setting the target saturation of each phase and the constraints of the period and phase duration,we construct a linear programming model with the minimum deviation between the actual saturation and the target saturation.Therfore,the algorithm of the signal timing optimization method is constructed.The SUMO simulation is used to obtain the license plate recognition data under five traffic demand levels.The simulation results which are optimized by the method proposed in paper and the Wester method are compared and analyzed.The simulation results show that the proposed method can reduce the total delay and the saturation of each stage is close to the setting target value.The research in the paper consider the current situation of signal timing optimization at isolated intersection,Time-of-Day method and signal timing optimization method based on license plate recognition data are researched,It will provide certain technical support for the coordinated optimization of urban road and regional traffic optimization.