Coupled Game Theory-kinematics Decision Making For Automatic Driving Lane Change

With the development of computer and information technology,autonomous vehicles has received more and more attention.Autonomous vehicle are important development direction of the automotive industry,and play a vital role in traffic safety and congestion.In recent years,autonomous driving technology has achieved certain results,but still has to face many difficulties in achieving safe and efficient driving in a real road environment.In the development process,autonomous driving will inevitably experience a hybrid driving environment to a fully autonomous driving environment.The lane change algorithm is the basic algorithm in the decision planning module of the autonomous vehicle,which is of great significance to driving safety.Therefore,the two driving environments should be separately discussed for the research on lane changing problem.The conflict between lane-changing vehicle and the rear vehicle on the target lane is an important issue in lane-changing research whether in the hybrid driving environment or in the fully automatic driving environment.At present,there is no relevant model for how to solve the conflict between the lane changing vehicle and the target lane during the lane change in the fully automatic driving environment.The research on this problem in the hybrid driving environment is not comprehensive enough.In this paper,the vehicle conflict problem in lane-changing process is taken as the research object and a game theory-kinematics coupling model method is proposed to study the decision-making problem of auto-driving vehicles in lane-changing process.Aiming at the two different driving environments of fully automatic driving and hybrid driving,two different automatic driving lane-changing models are established.The vehicle lane changing-avoidance decision rules are designed for lane-changing vehicles and the rear vehicle on the target lane in fully automatic driving environments,and the nested lane-changing game algorithm is designed for automatic driving in hybrid driving environments.The lane-changing models of automatic driving vehicles in two driving environments are established respectively.In fully automatic driving environment,the vehicle lane changing-avoidance decision rules are designed for lane-changing vehicles and the rear vehicle on the target lane.The rule first establishes the criteria for the occurrence conditions of lane-changing game;Secondly,the benefits of vehicles involved in lane-changing process are considered and the strategies used by the two cars and the corresponding longitudinal acceleration under each strategy are obtained by kinematics method.Finally,the two vehicles use their own strategies to avoid potential safety hazards in lane change conflicts.In the mixed driving environment,the paper first establishes a multi-step dynamic lane-change game framework based on strategy.Secondly,the benefits of vehicles involved in the process of automatic lane change are considered.Thirdly,it proposes an auto-driving nested lane-changing game algorithm to solve the strategy of lane-changing vehicles and the corresponding longitudinal acceleration under the strategy,while the rear vehicle on the target lane uses human vehicle strategy and acceleration model to carry out the strategy.Finally,the two vehicles use the multi-step dynamic lane-changing game framework to play the game until the conditions of lane-changing termination are satisfied.The results show that the model designed in this paper solves the conflict between the lane-changing vehicle and the rear vehicle on the target lane in the fully automatic lane-changing environment.In case of conflict,the two vehicles can choose their own strategies according to the vehicle lane change-avoidance decision rules designed in this paper,and the conflict between the two vehicles can be solved on the basis of ensuring safety and fairness.In the hybrid driving environment without satisfying the lane change conditions,the automatic driving vehicle can play multiple games with human vehicles by the nested game algorithm,and actively strive for lane change space to achieve a certain probability of lane change on the basis of ensuring the safety of both vehicles.In addition,this paper designs a strategy-based lane change game framework in a hybrid driving environment,which not only ensure the safety of the automatic driving vehicle in the lane-changing game with human vehicles,but also reflects the dynamic game process of the actual lane change.