Research On Hierarchical Control Algorithm Of Motion Cooperation For Vehicle Platoon Based On Game Theory

The cooperative control of vehicle platoons utilizes intelligent networking and intelligent driving technology to automatically adjust the spacing and yaw angle between vehicles to achieve a preset formation.With reasonable spacing,the platoon can effectively save energy and reduce the probability of traffic accidents.These potential advantages make the research on cooperative control of the vehicle platoon attract increasing attention.With vehicle automation development,vehicle platoons' studies have gradually expanded from onedimensional scenes to two-dimensional scenes.As a result,further problems to be solved have arisen.For example,the researches on behavioral decision-making are insufficient.The competition and cooperation relationship of the decision-making of the longitudinal distance between vehicles in the platoon are not considered,and the longitudinal and lateral motion coupling effect exists in the two-dimensional scene.It is challenging to deal with rights of way conflicts with vehicles outside the platoon,etc.In response to the above problems,this paper conducts a game-based vehicle platoon hierarchical control algorithm sponsored by ‘Dynamic Modeling and Cooperative Control of Distributed Electric Vehicles Based on Wire Control Chassis'(Science and technology project of Jilin Provincial Education Department,No.JJKH20200963KJ).By proposing a hierarchical control algorithm framework including vehicle platoon-level behavior decision-making,platoon-level motion control,and vehicle-level chassis control,complex multi-vehicle cooperative control is carried out with functional decomposition.The problem of right-of-way conflict inside and outside the platoon is abstracted as an interactive scenario of "cooperation in the platoon and competition outside the platoon," which aims to solve the problem of the right-of-way conflict and space allocation of vehicles in the multi-vehicle coordination process,and finally,realize the cooperative control of the vehicle platoon.Based on the above ideas,the main research contents of this article are as follows:(1)Aiming at the right-of-way conflict and coupling effects of vehicle platoons in twodimensional scenes,this paper proposes a hierarchical control framework,including a platoonlevel behavioral decision layer,a platoon-level motion,and vehicle-level chassis control layer.The framework includes the platoon-level behavior decision,the platoon-level motion control,and the vehicle-level chassis control.Through functional decomposition and hierarchical transfer of control flow,the complexity of the vehicle platoon's cooperative management is reduced;to meet the driving requirements of the two-dimensional scene,the motion control and control allocation functions at the vehicle level.In this paper,the reference vector field path following algorithm is used to construct the longitudinal spacing control and lateral path tracking.Combined with the Hamilton function,the H-RVF path tracking method is proposed;combined with the pseudo-inverse matrix and the Hamilton function,this paper develops a pseudo-inverse matrix method for control allocation with cycle and correction.The constrained control allocation is realized by circle correction with the aims of improving accuracy and real-time performance.(2)Considering the problem that the spacing strategy adopted by the current platoon cannot distribute the driving space equitably,this paper regards the spacing control strategy as space allocation.The bionic motion principles and cooperative games are combined to allocate space.This strategy considers collision avoidance,speed matching,center clustering,and other motion principles,as the follow-up cooperative game profit function,and provides a basis for space allocation.The interaction of the platoon vehicles is abstracted as a directed graph cooperative game problem,and the characteristic function that satisfies individual rationality and group rationality is proposed.The cooperative game solutions such as ? value and AL value are applied to allocate spacing,and the stability of the vehicle platoon control strategy is analyzed in combination with the longitudinal motion controller.(3)Aiming at the driving scenario where there is the conflict between the vehicle platoon and the vehicles outside the platoon,by analyzing the non-cooperative game structure of the platoon,this paper proposes a complete information game model for the unlimited platoon strategy and the limited strategy incomplete information suitable for the right of way competition with external vehicles Dynamic game model.Combining factors such as safety,self-driving tasks,economy,etc.,a non-cooperative game profit function is proposed;with acceleration as the action strategy,a complete information game model of infinite strategy is presented,and the extreme point is judged by evaluating the quasi-concave nature of the cost function Existing sufficient conditions and obtaining an equilibrium solution on this basis,comparing the performance difference between the cooperative cost function and the selfish cost function in the merge condition,the results show that the final benefit of the cooperative cost function is better than the selfish cost function;Construct a belief pool used to characterize the probability of the type distribution of both parties in the game,update the belief pool with driving intention identification and Bayesian probability formula,obtain the posterior probability of the type distribution,and obtain the conflicting two parties' values through Bayesian Nash equilibrium solution The stable behavior strategy analyzes the stability of the Bayesian Nash equilibrium to ensure that the conflicting parties will not actively change their driving behavior during the game.(4)Finally,based on Car Sim and MATLAB/Simulink,a vehicle platoon collaborative control simulation platform was established,and the layered control algorithm proposed in this paper was verified and analyzed.MATLAB is used to build a visual interface,including working condition setting interface,behavior decision interface,motion control interface,control allocation interface,and execution control interface.The interfaces can help users set control parameters and observe the driving performance.Three simulating scenes are constructed,including leading vehicle' acceleration and deceleration,double lane change,and the right-of-way conflict condition.The results show that in the acceleration and deceleration scenario,the common spacing strategy based on clustering behavior is based on The Sharpley value,? value,and AL value spacing strategies can all realize the stable longitudinal tracking of the platoon.Among them,the AL value is better than other methods in terms of error convergence and stable time during the tracking process.Besides,it shows a good consistency control effect,but this strategy requires all vehicles' status information in the platoon.In the two-dimensional scene,the vehicles in the platoon can realize the longitudinal and lateral coupling control through the H-RVF algorithm.The speed fluctuates good convergence characteristics and avoids the hidden safety hazards caused by the volatility amplification effect.In the right-of-way conflict conditions,the joint probability distribution in the belief pool converges to a stable state with the update,and the platoons and merged vehicles are implemented for each other's types.After making accurate judgments and behavioral decisions,vehicles formed a new platoon and quickly entered a new stable state under the condition of ensuring a safe distance.