Study Of Property On Traffic Flow State In Artery

Urban artery system takes in the main transportation. Traffic volume varies greatly and the congestion appears frequently in peak hours. This paper studies the property of the traffic flow state and gets the car-following property and recognizes the traffic flow state transformation of the artery section and the traffic flow characteristic of the intersection, which is very important on controlling the congestion, enhancing the transformation from the congestion to the scatter, carrying out the harmony control and increasing the capacity of the road network.The traffic flow state is classified as the congestion, the car-following and the scatter by the traffic flow state index. The classification standards of traffic flow states are provided by the Self-recognized Neural Network. The speed-density model parameters are fixed on by the regression, which is used to calculate the expect space headway. The Greenberg expect space headway model is set up based on the Greenberg model. The car-following model based on the expect space headway is established considering the relationship between the acceleration and the speed difference and the space headway. The model parameters are calibrated by the nonlinear regression function of Matlab. The acceleration non-symmetry property of the established car-following model is proved by analysis. The disturbing course of the acceleration and the deceleration are simulated by the numerical simulation. The acceleration sensitivity is put forward to describe the dynamic property of the traffic flow state. The definition of the acceleration sensitivity is provided and its expression is deduced by the Taylor spread based on the established car-following model. The regularity of the acceleration sensitivity variety in the traffic flow state transformation is analyzed and the rational value range of the quantitative analysis in the state transformation is suggested. On the base of the hypothesis of the density continuum, the improved starting weave model and the parking weave model are established. The calculation expressions of the queue length and the scatter time at intersection are provided. The effect of the road condition, control type and the mixed traffic flow on the traffic flow state is analyzed and the advice is provided for the drivers and the managers to make sure the traffic flow fluent.The main achievement of this paper consists of the following items. Traffic flow state is classified to three states by Self-organized Neural Network; the expect space headway model and the car-following model based on the expect space headway are set up. And then the acceleration non-symmetry property is proved. The course of the acceleration and deceleration disturbance on the traffic flow is analyzed. The concept of the acceleration sensitivity is come up with and its expression is deduced. The traffic flow transformation regularity is analyzed and the rational span of the quantitative analysis in the state transformation is suggested. The improved starting weave model and the parking weave model are established. The calculation expressions of the queue length and the scatter time at intersection are provided.