Game Analysis And Simulation Research On Vehicle Lane Changing Behavior On Urban Roads

Vehicle lane change is one of the most common traffic behaviors.After the driver has the intention to change lanes,he will go through two processes:decision-making and lane change execution.Decision-making judgment is a process of communication and game with the driver behind him that is driving on the target lane,this determines whether lane changing can be implemented.Therefore,these two steps have quite a lot impact on the traffic flow.However,the current research on lane change decision-making and lane change implementation is mostly to separate the two and carry out research independently,a lane changing simulation model that can reflect information interaction is needed.Based on this,this paper firstly carries out an investigation and collect lane changing behavior data,then sorts out macro traffic data such as traffic flow,lane changing times,lane changing methods,etc.With the video processing software,the microscopic characteristic data of a single lane-changing vehicle are extracted,including the lane-changing position of the vehicle,the driving speed and variation before and after lane-changing,the distance between the vehicle and the behind vehicle which is on the target lane,lane-changing time,etc.SPSS and Origin are used to process the data statistically and test the normal characteristics of the data.Secondly,the paper establishes a concise model of vehicle lane changing decision game.The model takes the lane change vehicle CV and the vehicle running on the target lane that is behind him FV as the main body of the game.It explains the results and income under various decisions of CV and FV,also calculates the dynamic evolution equation of the proportion of drivers who adopt the two strategies of lane changing and concession.All stable solutions of the model are listed one by one,and stability analysis is carried out by combining with the evolution path map drawn by MATLAB.The concise model is subdivided into free lane-changing game and mandatory lane-changing game.From the three angles of speed,safety and lane-changing demand,the abstract income is concretely expressed by combining the data of velocity variation,distance between CV and FV and lane-changing position sorted out in chapter 3.Maximum likelihood estimation method is used to calibrate the parameters in the two models.The images of lane changing probability and comity probability after parameter calibration is drawn by MATLAB,which are compared with the actual survey data to prove the accuracy of calibration results.Finally,the paper establishes a simulation model describing vehicle lane changing process.The model is based on STNS model to improve its strict lane change rules.The model is also divided into free lane-changing and mandatory lane-changing according to lane changing intention.The key point of model improvement is that it considers not only the lane changing behavior of CV,but also the game process between CV and FV by introducing the comity probability of FV vehicles.CV lane changing probability and FV comity probability are both controlled by the dynamic equation presented in this paper When the distance between CV and FV is between the minimum lane changing distance and the safe distance,whether lane changing behavior is allowed or not is judged according to the comity probability,and then the model is updated according to the rules of NaSch.After calibrating various parameters in the model,taking the forced lane change model as an example,the investigation section is simulated,and the validity of the model is verified according to the output flow and lane changing times.Compared with STNS model,it is proved that this model can weaken the phase separation phenomenon and is more suitable for describing the actual traffic situation.Therefore,traffic suggestions can be put forward after analyzing the model parameters,also this model can be used for related research on the influence of lane change behavior on traffic flow.