Study On Ramp's Traffic Operation Mechanism Of Urban Expressway And Design Of Ramp Control

With the development of our country's economic and Society and the accelerated process of urbanization, urban expressways network in many big cities of our country have enjoyed rapid development, it is ensure to travel within the city rapid, and convenient to play an extremely important role. However, owning to the rapid growth of traffic demand, the construction speed of urban expressways is slower than traffic growth, coupled with China's lack of holistic research of traffic flow's characteristics and control technology of urban expressways, a large number of traffic get into main lines out of order from the on-ramp, which seriously affect the natural exertion of expressways and traffic safety and traffic jam of expressways is getting worse. So it is necessary to research urban expressway ramp traffic flow characteristics and control technical.Based on the measured data, this paper analyzed in-depth the on-ramp traffic flow and the impact of the main line of traffic flow, in accordance with the Ramp traffic flow characteristics proposed short-term traffic-flow forecasts for the designation, Established the model of on-ramp short-term traffic flow forecasting based on the multi-model neural network, proposed a on-ramp control strategy mainly to the main line, take account into the changes of traffic demand, and designed the System's framework, process control function of each module in accordance with the control strategy, then use PTV/Vissim software simulation of the control system to evaluate the reasonableness of the system design.Through analysis of Shanghai urban expressway test data showed that, on-ramp traffic flow of urban expressway has seen an obvious"periodicity","pulse"and"fluctuation", and such regularity caused the disturbance of main line traffic flow patterns.Analysis showed that the measured, according to the pulse cycle time of on-ramp traffic flow to set the forecast long can reduced the ramp traffic flow's volatility of the time series, and improved forecasting accuracy. In establishing the method based on multi-model neural network greatly improving the flow of short-term traffic forecast accuracy and stability of forecasts.The simulation results show that, the adjustment model proposed in this paper can adjust the next rate adjustment according to the current ramp traffic demand and the size of the margin rate adjustment real-time, the control effect surpasses the fixed time control method obviously. Traffic during peak hours or when major changes in the situation of control effect more pronounced.