Research On The Adjustment Of Signal Timing Based On High-precision Trajectory Data

Scientific setting and optimization of signal light timing is one of the indispensable conditions for improving the traffic capacity of signalized intersections,alleviating urban traffic congestion,and building a smart city.At present,the new traffic detection technology has overcome the limitations of traditional cross-section traffic flow detection technology,making it possible to obtain real-time and accurate intersection full-sample vehicle trajectory data,and providing new opportunities for realizing more accurate signal timing optimization.In view of this,facing the actual needs of the optimization of singlepoint signal timing schemes on urban roads in my country,this article proposes a twolevel fuzzy control strategy for signal lights based on genetic algorithm to achieve datadriven online closed-loop real-time optimization of signal timing parameters under different traffic conditions.The specific research content and results are summarized as follows.First,this article preprocesses the vehicle’s high-precision trajectory data from five aspects: data cleaning,data correction,trajectory reconstruction,data thinning,and format conversion.The key trajectory data is extracted by setting the event criterion and virtual coil.Secondly,based on the pre-processed high-precision trajectory data,the vehicle running characteristics are analyzed.According to the analysis results,multiple parameters are calibrated such as the expected speed distribution,acceleration distribution,deceleration distribution in the VISSIM simulation model,and the parking distance,additional part and the multiple part of the safety distance in the Wiedemann 74 model.Then,the validity of the calibration parameters was tested according to the measured data,laying a foundation for the establishment of a micro-simulation platform.Then,in view of the actual situation of Beier Road and Chongzhi Avenue,based on the traditional two-level fuzzy control strategy of traffic lights,the article established a complete set of two-level fuzzy control rules for traffic lights,and use genetic algorithm to continuously optimize the membership function of input and output variables in fuzzy control.Making the timing strategy have the ability of autonomous learning,and to ensure that the timing strategy still has an excellent control effect when the traffic flow state of the intersection changes significantly.Finally,this article takes the intersection of Beier Road and Chongzhi Avenue in Shenzhen as an example,and takes the traffic volume and composition data of the intersections in different periods as the input of the micro-simulation platform to simulate the traffic operation conditions of the intersections between 5 o ’clock and 20 o ’clock on March 13 th.Based on the simulation results,the rationality and stability of the signal timing strategy proposed in the article are comprehensively evaluated.