Research On Data Delivery Scheme Based On Fuzzy-rule And Machine Learning In Vehicular Networks

Data delivery in Vehicular Networks(VANETs)is a challenging task due to the high nodal mobility and constant topological changes.In common routing protocols,multi-hop V2 V communications suffer from higher network delay and lower packet delivery ratio(PDR),in addition,search and interrupt reconstruction of communication links will increase network overhead.In the process of urban traffic operation,a large amount of traffic data will be generated continuously,which have an important guiding significance for the further research on data delivery in VANETs.In this paper,we propose a novel data delivery scheme for vehicular networks in urban environments,which takes full advantage of these dense traffic data to improve data transmission quality(DTQ).A fuzzy-rule-based data transmission approach is designed to optimize V2 V communications,the vehicle-based short-term speed prediction method makes the data communication links more reliable,and the dynamic predictions of the vehicles are realized with the assist of specially designed machine learning system(MLS).The main contents of this paper are as follows:(1)A fuzzy-rule-based data transmission approach is designed.This approach can optimize relay node selection by considering multiple factors comprehensively,including vehicle speed,driving direction,hop count,and connection time.By optimizing wireless communication between V2V(vehicle to vehicle)or V2R(vehicle to RSU),DTQ will be improved.Paths are evaluated according to the fuzzy comprehensive evaluation method(FCEM),wherein the index weight vector is determined by the analytic hierarchy process(AHP).This method makes the selected data transmission paths more stable and reduces the losses caused by the link interruption and reconstruction,not only guarantees fast and efficient message dissemination,but also alleviates broadcast storm.(2)The weighted KNN based short-term vehicle speed prediction is proposed.Different from the normal segment-based speed prediction method,this paper introduces vehicle-based prediction method.By predicting vehicle speed at next moment,the selected relay link will be more reliable.In order to improve the prediction accuracy and reduce the computational cost,this method systematically considers spatial and temporal influence factors.We select four variables that playing major roles for dynamic speed prediction,that are velocity,acceleration,the vehicle count in the testing sample's communication range,and the vehicle count gradient.(3)An assisted decision making based RSU machine learning system(MLS)is designed.The KNN algorithm based MLS is embedded in each RSU to send packets to the destination vehicle in the blind zone.Real-time traffic data is collected by the mutual cooperation between RSUs and vehicles,and then uploaded to the MLS.The MLS will analysis and process the received latest data in real time based on historical traffic data,and provide dynamic routing decisions to the data transmission system ultimately.This method will first predict the turning direction at the intersection for the destination vehicle,and then predict the next RSU that vehicle will enter,finally acquire the rough location of the target vehicle.Finally,the performance of our proposal has been verified through simulations,and different simulation parameters are set to simulate the data transmission under different traffic conditions.The results show that,compared with contrastive schemes,our proposed data delivery scheme can improve the packet delivery ratio,guarantee the timeliness of messages,and reduce the control overhead significantly.