BP Neural Network Based External Left Turn Lane Delay Prediction And Setting Decision

With the rapid development of economy in our country,the city of motor vehicle ownership is increasing,the development of the road system to keep up with the development of traffic demand,road signal intersection has become the bottleneck of the road network traffic capacity and driving safety,and set up the left car lanes is an effective way to improve the performance of the intersection,and properly set the left lane is is the key.The traditional builtin left-turn lane has some defects such as small turning radius and obvious interleaving phenomenon.However,in recent years,some cities have adopted the outer left-turn lane one after another,which has reduced the interweaving phenomenon of left-turn vehicles and direct traffic in the upstream of the entrance road at the intersection and reduced the traffic delay.Some scholars also believe that this kind of traffic organization,which is contrary to driving habits,will lead to some careless drivers not to follow the lane,which will have a negative impact on the overall operation quality of the intersection.Based on the current research,there is no clear boundary between the setting position of the left-turn lane.Aiming at the above problems,with the goal of minimizing traffic delay,the influencing factors and determination methods of left-turn lane outboard are studied.Using VISSIM traffic simulation software,the simulation models of different number of straight lanes,interleaving length and interior and exterior position of left-turn lanes were constructed and the parameters were calibrated according to the measured traffic data.Through the simulation analysis of the number of vehicles in different sections and the number of left transfer vehicles in the outer lane,the delay data under the above different geometry and traffic conditions are obtained.MATLAB is used to analyze the relationship between the length of the interleaved section and the number of left-turn vehicles on the traffic delay of the left-turn lane in and out of the lane under the condition of different vehicle Numbers on the straight line,and the change rules are discussed.The BP neural network model was constructed,and the network training was carried out according to the number of straight lanes,the length of interleaved segments,the number of straight lanes and the number of left-turn vehicles.When the training error was less than the ideal error,the driving delay of the in-left lane and the out-left lane was obtained.Through the delay comparison,select the condition with less delay as the setting position of the left-turn lane.The traffic parameters of the four intersections were investigated.The BP neural network model is used to analyze and calculate the traffic delay and determine the internal or external scheme of the left-turn lane.Verify the decision result through VISSIM delay analysis.The results show that the more left-turn vehicles in the outer lane,the more vehicles in the straight section,the shorter the length of the interleaved section,and the more straight lanes,the more obvious the delay reduction effect of the left-turn lane is.The BP neural network model can be used to determine the setting position of the left-turn lane through delay analysis,and the result is in good agreement with the simulation calculation and the practical scheme.