Research On Key Technologies Of Active Driving Service System For Internet Of Vehicles

With the rapid development of intelligent,networked,electric and shared "new four modernizations" technology,driving service system has gradually become an important research direction in the field of intelligent vehicles and intelligent transportation.In recent years,the driving active service system(DASS),which can actively and effectively perceive and identify the service needs of driving users,and can actively push services to driving users and provide effective guarantee for service quality,has become a research hotspot.As the executive carrier of driving process,intelligent vehicle can grasp driving information more fully,improve driving safety through service platform,and have important application value for promoting the development of driverless technology.The main research contents are as follows:(1)In order to meet the requirements of DASS multi equipment multi vehicle collaborative reasoning,information processing and hierarchical centralized control,based on the multi-agent collaborative control framework,the hierarchical centralized collaborative DASS architecture and service process are constructed.The perception and decision-making ability of each level in DASS lays a foundation for the design of DASS service process and its key algorithm.(2)In view of the traditional driving service methods can not meet the needs of automation,a service generation platform based on multi-agent is proposed.Through the process of driving environment modeling,service demand modeling,executable capability matching and service evaluation,DASS has the capability of demand matching and service generation.In the mixed traffic simulation scenario,on the one hand,the accuracy and timeliness of service generation is verified by the service matching degree;on the other hand,the initiative of service generation algorithm is verified by the continuous traffic flow characteristics of different strategies in specific driving scenarios.(3)The active push platform of DASS and its main algorithm are constructed and verified by simulation.Aiming at DASS neglecting the interaction between service push and driving users,a driving push service platform based on user characteristic behavior is designed,which includes storage layer,application layer and evaluation layer.In the application layer,B-Num.BT algorithm is designed based on the operation behavior prediction of driving users,which improves the service push user acceptance.In the evaluation layer,the evaluation index of push performance is designed based on push process.Finally,the timing,safety and accuracy of typical push scene are verified and tested by PreScan/Simulink,and the relevant timing is discussed to classify the driving users.(4)According to the demand of active control in the sudden scene,the intelligent vehicle avoidance control model is designed for the dangerous parallel scene,and the performance of active intervention control in the sudden scene is verified by DASS.Firstly,BP neural network is used to simulate the driver's emergency avoidance behavior,and Gaussian mixture hidden Markov probability factor is added to predict the merging probability of side vehicles.Finally,the applicability of the avoidance model to the scene of dangerous Lane merging is verified by the speed difference experiment and the penetration experiment of connected vehicles,and the service ability of DASS is verified by the timing of intervention.In conclusion,this paper further breaks through and improves the intelligent,personalized and acceptable performance of DASS from the model,algorithm,platform and application level,forming a more detailed and feasible DASS implementation scheme,which further enriches the theoretical methods and application approaches of intelligent transportation.