Research On The Theory And Method Of Multi-ramp Coordinated Control Of Urban Expressways

As the social economy develop rapidly,the private ownership of small cars keeps growing,and the demand for highway traffic is constantly increasing,and then higher requirements are put forward for the management of expressway.The coordinated control of freeway entrance ramp provides relief measures for the highway congestion.Considering the existing of the deficiency of control scheme based on traffic prediction,as well as the lack of road network abnormal condition monitoring scheme in the process of ramp control scheme,the traffic state judgment ramp coordinated control scheme is proposed in this paper.This research mainly does the following work.(1)In the research of basic data prepare of multi-ramp control,a short-term traffic flow prediction scheme based on support vector machine regression is proposed and implemented.The forecast is divided into two parts.The short-term traffic flow forecast provides the forecast flow data for the control scheme.Short term traffic flow average velocity prediction combined with traffic flow prediction is used to judge congestion states of main line.Based on the background of the large vehicles had a greater influence on the average speed in mountainous area,using the simulation,and the theory of gray relational method to analyze the influence degree of large vehicles scale for the average speed,in the short-term traffic flow prediction cart factors predicting scheme was proposed,and through the optimal input time series analysis,comparative analysis of the neural network prediction model for the optimal prediction model.The optimal scheme prediction results are as follows: the short-time traffic flow velocity forecast mean square error is 0.02419,and the determination coefficient is 0.58;Short-term traffic flow prediction scheme,the mean square error is 0.038384,and the determination coefficient is 0.71954.(2)In view of the lack of network anomaly monitoring scheme in the implementation of the current control scheme,which can’t guarantee the poor effect of the scheme because of abnormal events,an abnormal event detection model based on support vector machine classification is proposed.Through the simulation of the vehicle anchor event,the relevant parameters were collected and the detection scheme was designed.The single-side input detection rate was 91%,the false alarm rate was 4.5%,and the average detection time was 114 s.The double-side input detection rate is 93%,the false alarm rate is 3.5%,and the average detection time is 102 s.It is sufficient to show that the detection scheme designed in this paper has a better comprehensive effect.The two-side input scheme is slightly better than the single-side input,and the single side input can also achieve the detection effect.(3)Finally,the simulation model of Kun Yu expressway was established by VISSIM simulation software.And no control,improved ALINEA control,traffic state judgment single ramp control,ramp coordinated control four schemes are compared.The results show that the proposed single ramp control scheme based on traffic state judgment is better than the improved ALINEA control algorithm,which can alleviate the bottleneck area congestion,and the overregulation is small.The coordinated control scheme is more obvious than the single point control effect,the density parameter of the stable bottleneck area can be achieved,and the overall condition of the road network reaches the expected density.This paper establish coordinated control scheme based on traffic state judgment,under the condition of the guarantee network without accident,road network optimization is obtained by quantity,prevent the on-ramp queue overflow,It is suitable to prevent and relieve the bad traffic condition in the bottleneck area.