Research On Driver's Lane-Change Behavior Based On Evolutionary Game Theory

Recently,traffic safety accidents caused by lane-changing behaviors and reduced road operation efficiency have frequently occurred.The lane-change behavior and the following behavior are the two basic behaviors that constitute the vehicle running on the road.The research on lane-change behavior,especially the knowledge of its evolution in time and space,plays a key role in reducing road traffic conflicts and improving traffic efficiency.This paper did the following research work:Firstly,the two sides of the game and the income matrix of the change behavior were analyzed,and the rules of the lane-changing behavior on the two-lane road were formulated.Secondly,the lane-changing behavior simulation module based on NetLogo was established to build the simulation environment of roads and vehicles,and to give the behavior rules of vehicle acceleration,deceleration,speed maintenance and lane-changing behavior.Next,Simulation and analysis of the evolution of vehicle lane-changing behavior under different lane-changing behaviorsFinally,by designing the experiment based on vantage edge2,the evolution behavior of vehicle lane-change in the simulation stage was tested to analyze the effectiveness of the lane-changing behavior simulation module.The research in this paper has the following innovations:Firstly,combining the incomplete rational factors of the driver in the lane-changing behavior and the cooperative driving behavior that may occur in the lane-changing driving behavior,using evolutionary game theory,kin selection theory,reciprocal altruistic rules to analyze the lane-changing driving behavior and establishing the lane-changing behavior simulation module makes the model better applicable to reality.The second is to propose the concept of the correction factor based on the degree of civilization of the country and society and the credit index of the driver.At the same time,it is introduced into the game matrix of lane-changing behavior to realize the modification of the lane-changing behavior and realize the degree of civilization in different regions.The game matrix transformation of the driver's reputation index makes the model scalable.