Adaptive Traffic Signal Control Method Based On Dynamic Platoon Dispersion Model

Most of the input data for existing traffic signal control systems are collected from fixed detectors before the appearance of big data.The predicted traffic flow characteristics often differ greatly from reality due to the flaws in this kind of input data(limitation,dispersion,isolation,and hysteresis).As a result,the traffic signal control system cannot effectively respond to the changing traffic flow even when equipped with a highly intelligent traffic signal control method.This bottleneck has blocked the efficiency improvement in traditional traffic signal optimization method.With the development of connected vehicle(CV)technology,collection of massive real-time vehicle trajectory data becomes possible.When combined with the feedback data of traffic signal control system(vehicle delay,queue length,and emission),this opens a new way for the development of intelligent traffic signal control system.In order to meet the requirement of future transportation management and control,this study presents a dynamic traffic flow model based on connected vehicle data.Then,an adaptive control model and algorithm for urban traffic signals is constructed based on the proposed dynamic traffic flow model.Hopefully,the studies performed in this dissertation can lay a foundation for the establishment of a new generation of intelligent traffic signal timing optimization theory,which is in a CV environment satisfying the future traffic management and control requirements.Specifically,the research work of this dissertation mainly includes the following several aspects:(1)A micro-simulation environment for CV is built to verify the models and algorithms in this study.First,the setting and t echnical characteristics of CV environment are analyzed.Then,the requirements of traffic signal control for CV technology are obtained.Finally,a CV micro-simulation platform is developed based on Vissim and MATLAB.Specifically,Vissim and MATLAB are u sed to simulate traffic behavior and realize models and algorithms,respectively.(2)The dynamic platoon dispersion models are proposed based on travel time and travel speed,which obtained from CV.Firstly,a dynamic platoon dispersion model is developed based on the classical Robertson's model.The CV data are us ed to dynamically estimate the parameters of the proposed model.Then,a dynamic heterogeneous platoon dispersion model is proposed in a CV environment to capture the heterogeneity of traffic flow.Compared with the existing static model,these dynamic models can better reflect the dynamic characteristics of traffic flow.(3)A traffic signal timing method for isolated intersection is proposed based on dynamic platoon dispersion models.Specifical ly,different traffic signal timing methods are proposed for different prediction time interval of dynamic platoon dispersion models.For relatively short prediction time interval,a signal timing optimization model based on Stage/Barrier is developed and solved by dynamic programming.This method optimizes the traf fic signal plan in real time based on the arrivals predicted by dynamic platoon dispersion models.For shorter prediction interval,a signal timing optimization model based on a dynamic cycle tim e is proposed.A genetic algorithm is adopted to dynamically optimize the signal timing plan.Compared with traditional signal control method,these two methods can effectively reduce and balance vehicle delays in all directions of intersection.(4)A coordinated control optimization method for multiple intersection s is proposed based on dynamic platoon dispersion models to coordinate the control plan of intersections and road sections.First,the relevant constraints in coordinated control are analyzed.Then,a synchronous dynamic optimization model for intersection and road section is developed.In the intersection level,based on the above single intersection optimization control method,a new model is constructed considering overlap phase,and solved by a new algorithm based on rolling optimization strategy.In the road section level,based on dynamic platoon dispersion models,the delay utility function of road segment is analyzed,and a dynamic offset optimization model for road section is developed.Finally,the proposed dynamic coordinated control optimization method is validated by comparing with coordinated actuated signal control method.The above research results are beneficial to the exploration of urban traffic signal adaptive control system in a CV environment,the promotion of intellectualization of urban transportation control system,and lay a theoretical foundation for future urban traffic signal control.