| In recent years,with the gradual expansion of the city scale and the continuous improvement of social economic level,the development of freeway mileage become rapid.However,as a result of the freeway scale expanding and the traffic trips increasing,traffic accidents in freeway are becoming more and more frequent,which brings great loss to social economy.Real-time monitoring and accurately control freeway traffic flow can effectively protect the traffic safety and efficiency in effect,but the freeway detection technology in current,economic costs and other constraints cannot guarantee that the entire road network layout vehicle detector in high density.As the distribution of vehicle detectors is not dense enough,the comprehensive traffic state of the freeway cannot be known accurately.Therefore,this paper regards the continuous sections of freeway basic part as the research object.The measured data of traffic flow data parameters are collected to analyze their temporal and spatial distribution of continuous sections.Then the multi-lane macroscopic model of traffic flow is used to estimate the realtime traffic flow state of all lanes in the sections which don’t equip detectors,so that provide the theoretical method and technical support for the formulation of freeway management strategy.First of all,the measured data of traffic flow parameters are collected to analyze their temporal and spatial distribution of continuous sections in freeway basic part,and the appropriate state characterization parameters are selected to establish the data foundation for the traffic flow estimation.Secondly,by comparing the results of the traditional single-lane traffic flow model and the multi-lane traffic model for the traffic flow state estimation,Laval-Daganzo model is selected as the basic model for the traffic flow state estimation,and applied to the specific part of freeway.Then the parameters of the model are calibrated on-line by genetic algorithm.Thirdly,the model is applied to freeway for the traffic flow state estimate in different ways of detector distribution after parameter calibration,and the result will be evaluated and discussed,including the tracking ability of the traffic flow state fluctuation and the analysis of result comparison in different ways,so that reference for the method of distributing detectors is going to be provided. |
