Research on the Method of Setting Variable Lanes at Urban Road Intersections

In order to alleviate the current urban traffic congestion and ensure the safety of driving, according to the uncertainty and random fluctuation of road traffic flow data, an BP neural network algorithm and Webster model are proposed to optimize urban roads and improve vehicle traffic effectiveness. The prediction accuracy of the neural network algorithm is high, and based on the predicted traffic flow, the guide lane is changed before the actual traffic flow, provide a reference for vehicle driving and improve the utilization of road space. The design is based on the changed data of the lane attributes, using the improved Webster model to calculate the signal timing parameters and combining the specific conditions of the current road to optimize the traffic light timing and improve the time utilization of the intersection. To a certain extent, the space–time utilization of traffic roads has been improved, vehicle delays have been reduced, and traffic congestion has been alleviated. The traffic flow data at the intersection of Jinzhai Road and Tangkou Road in Hefei was used to test the proposed method, which verified the timeliness of the method.