Vehicle Path Planning Behavior of Variable Direction Lane Intersections Based on Game Theory

With the differentiation of urban functional areas, the problem of unbalanced intersection steering has intensified, and the imbalance of direction has gradually appeared. Existing studies generally use variable guidance lanes to solve this problem, but in practical applications, it is found that the utilization rate of variable guidance lanes is generally low. Under the background of vehicle-road collaboration, the emergence of CAV provides a new way to solve this problem. This study considered using the communication advantages of CAV to obtain the information of the lane ahead in advance, select the lane with low utilization rate, and plan the path ahead. Aiming at the behavior of vehicle prediction and path planning, this study adopts game theory to model and analyze, takes vehicle path as game strategy, uses B-spline curve to simulate longitudinal movement, uses sine function curve to simulate transverse movement, considers the scene characteristics, combines the safety efficiency of road section and the queuing of inlet road to construct the income function, and finally solves the model through the scribing method. Finally, SUMO (TraCI) and Python co-simulation method are used to simulate the imbalance of typical urban road flow direction. The experimental results show that CAV can indeed take advantage of its communication advantages to improve the utilization rate of variable guidance lane by way of advance routing decision. With the increase of CAV penetration rate, the intersection efficiency is significantly improved. The effectiveness of the proposed model is proved, and it can provide reference for the control of variable guidance lane under future vehicle-road coordination.