Lane-changing Behavior Of Intelligent Connected Vehicles On Accident-prone Highway Sections Based On Game Theory

Intelligent Connected Vehicle(ICV)is one of the most important aspects in the development of intelligent transportation,covering areas include internet of vehicles,Cooperative Vehicle Infrastructure,and autonomous driving,as such,ICVs offer both a promising market potential and significant research value.Lane changings is one of the most common traffic behaviors but the process is easily affected by various factors such as surrounding vehicles and driving environment,which could lead to traffic accidents.Intelligent connected vehicle enables multi-channel information exchange,supplying advanced driver assistance systems with extensive data concerning the surrounding environment and nearby vehicles.This informational support fostered effective decision-making and held promise as a proficient solution for mitigating traffic congestion and minimizing traffic accidents in the future.In particular,for highrisk road sections like highways,vehicles that were traveling upstream of an accident on the roadway should take appropriate measures,such as changing lanes or reducing speed,to avoid collision and ensure driving safety.Vehicle lane-changing behavior encompasses a process of competition and cooperation between vehicles,and can be regarded as a type of game.As a result,game theory-based intelligent connected vehicle lane-changing models are progressively emerged as a focal point in research.Nevertheless,there was currently limited research on the lane-changing behavior of vehicles after traffic accidents occur on highways.Consequently,this study endeavors to examine the manner in which intelligent connected vehicles employ real-time information to execute secure and effective lane changes following traffic accidents on highways.Firstly,the lane-changing behavior of vehicles was analyzed by considering the characteristics of such behavior.The study provided a summary of factors that influence lane-changing behavior,and proposed benefit functions based on speed and safety considerations.A vehicle lane-changing model based on game theory was constructed by considering specific lane-changing scenarios of the vehicle.Secondly,the study elaborated and summarized theories related to the spatiotemporal impact analysis of highway traffic accidents,including Traffic Wave Theory and the duration of traffic accidents.The Greenshields velocity-density relationship model is also used to simulate the operational state of traffic flow after an accident on the basic section of the motorway,and by analyzing the queuing wave and rarefaction wave after traffic accidents,the scope of traffic accident impact can be obtained.Thirdly,the interaction behavior of vehicles in the area affected by a highway accident is analyzed based on game theory,and the game payoffs is calculated using safety payoff and speed payoff functions.Combining the latest lane-changing point and the lane-changing pressure,a comprehensive lane-changing model based on game theory,which combines free lane-changing and forced lane-changing,is established for the region after accidents occur on the basic section of the highway.Finally,the study utilizes the SUMO simulation software and lane-changing data extracted and denoised from the NGSIM dataset to simulate vehicle driving behavior under different conditions.Comparative experiments are designed to validate the effectiveness of the lane-changing decision-making approach proposed in this study.Simulation results indicate that the comprehensive lane-changing decision model based on game theory not only ensures the safety of lane-changing behavior but also improves the average vehicle speed when passing through accident-prone sections.This model can assist vehicles in completing lane changes while ensuring both safety and efficiency.