Research On Key Technologies Of Horn Intention Recognition In Mixed Traffic Environment Of Intelligent Connected Vehicles

In recent years,with the rapid development of Intelligent Connected Vehicles(ICV),ICV is pouring into the flow of Manual Vehicle(MV),and the mixed driving situation is gradually taking shape.The interaction intention recognition between heterogeneous vehicles in the mixed traffic environment has become a research hotspot in the field of intelligent transportation.How to accurately and quickly identify the whistle of MV in ICV has become a difficulty in identifying the interaction intention of heterogeneous vehicles in mixed traffic flow.At present,scholars have not formed a complete and effective methodology for the research of the interaction intention of heterogeneous vehicles,and often ignore the auditory information in the driving environment.In this paper,the Perception-Location-Recognition(PLR)model and the Mixed Vehicle Fog(MVF)are established for ICV whistle intention under the strict requirements of accuracy and timeliness security.In this paper,the Clustering Fog Algorithm(CF)for ICV response to whistle event is proposed.The event related ICV is clustered and the underlying communication architecture for collaborative whistle intention recognition is constructed.The Fault-tolerant Node Clustering Resource Scheduling Algorithm(FNC-RSA)is designed to build the underlying computing architecture of collaborative whistle intention recognition.The main research work of this paper is as follows:(1)In view of the demand of perception of MV’s whistle intention in ICV mixed traffic environment,this paper analyzes the whistle characteristics from three aspects of "perception location recognition",and establishes PLR model of ICV’s whistle intention by combining Deep Convolution Recurrent Neural Network Algorithm,Time Difference of Arrival Method and Support Vector Machine Algorithm based on Motion Time Window.The results show that the average accuracy rate of whistle perception is 90.4%,the estimation error of location angle is less than 5 °,and the recognition rate is 82.5%,which basically meets the needs of ICV horn intention recognition.(2)In view of the lack of communication ability in the bottom layer of directional data acquisition and transmission of PLR model,MVF model is established and Clustering Fog Algorithm(CF)is proposed to integrate ICV responding to horn into a vehicle fog,which provides a reliable local communication network for task unloading and resource complementation in the fog.The results show that different ICV permeability has an impact on the event coverage of the model.Under the high permeability of 70% ~ 90%,MVF model has obvious effect.It can keep four members in each cluster,and keep a relatively fixed number of fog groups in the road network,which can effectively cover all whistle events and realize the integration of road network information.(3)In view of how the MVF model completes the task unloading and resource reasonable allocation of horn event calculation,The Fault-tolerant Node Clustering Resource Scheduling Algorithm(FNC-RSA)is proposed.The results show that FNC-RSA algorithm reduces the average hops by 55.17% and the average task completion time by 45.40% compared with the Low Energy Adaptive Clustering Hierarchy.Under the delay interference,FNC-RSA algorithm still maintains the high stability of mixed traffic flow and the safety of driving effectiveness,which provides the underlying computing architecture support for resource allocation and computing load balancing of PLR model.To sum up,aiming at the problem of ICV identifying MV horn interaction intention in mixed traffic flow,from the micro level,this paper proposes PLR model as the basic algorithm of MV horn intention recognition,and realizes the accurate recognition of horn intention through "perception location recognition".From the middle and macro level,MVF model is proposed as the bottom communication computing architecture of PLR model.Through the establishment of collaborative computing fog group in response to horn event,the timeliness of PLR model in the complex computing process is made up.Starting from the micro problems,the paper designs a micro medium macro comprehensive solution,which provides a new solution for the heterogeneous vehicle conflict under the mixed traffic flow.