Research On Key Technologies For Enabling Strong-timeliness And High-Reliability Communications In Vehicular Networks

The widespread use of vehicles in recent years have escalated issues such as road safety,traffic congestion and environmental pollution.Vehicular networks provide a promising solution to mitigate these problems.Thereby,real-time information interaction via vehicleto-vehicle(V2V),vehicle-to-infrastructure(V2I)and even vehicle-to-everything(V2X)communications can be enabled by vehicular networks to not only alert drivers(human drivers or vehicle controllers)about potential hazards and improve driving safety,but also assist driving planning,improve traffic efficiency,promote energy conservation and emission reduction.However,different from traditional wireless networks,vehicular communications have witnessed unprecedented challenges.For example,the network topology of vehicular networks changes highly dynamically in time and space and,at the same time,the diverse traffic environments and road topologies lead to a complex wireless signal propagation environment.Moreover,vehicular applications such as road safety,autonomous driving and traffic efficiency usually require stronger timeliness and higher reliability in information and communication.To tackle these challenges,this dissertation investigates key technologies for enabling strong-timeliness and high-reliability vehicular network architecture,V2 V communications,V2 I communications,and V2V/V2 I hybrid communications.Considering that data packet transmissions for different vehicular applications may exhibit different degrees of couplings in time,we have proposed to characterize the information timeliness in decoupled/weakly coupled and strongly coupled data packet transmissions using the data packet transmission latency and the age of information(Ao I),respectively;the latter is defined as the time elapsed since the freshest received packet was generated till the current time.The contributions of this dissertation are as follows:Firstly,a novel programmable architecture of fifth generation(5G)software defined vehicular networks which integrates the 5G mobile communication technologies,cloud computing,fog computing and software defined networking is proposed to achieve realtime,flexible and efficient management control of vehicular networks,and to promote strong-timeliness and high-reliability vehicular communications.Furthermore,the fog cell structure is proposed at the edge of 5G software defined vehicular networks to avoid frequent handovers between the vehicles and roadside units.Finally,the communication performance such as the data packet transmission latency in the fog cell is analyzed.Secondly,by modeling millimeter-wave(mm Wave)multi-hop V2 V communications in urban grid road scenario based on a novel framework of stochastic taxicab geometry,the data packet transmission latency and reliability under mm Wave multi-hop V2 V communications are investigated.A novel V2 V relay selection algorithm based on joint optimization of the forward progress and latency is proposed and compared with three existing V2 V relay selection algorithms,i.e.,random with forward progress(RFP),most forward with fixed radius(MFR),and nearest with forward progress(NFP),in terms of multi-hop transmission delay and reliability.Simulation results show that the proposed V2 V relay selection algorithm outperforms the RFP,MFR and NFP in both multi-hop transmission latency and reliability.Thirdly,the quality of service(Qo S)performance of vehicular networks for V2 I communication based vehicular real-time status monitoring is investigated,taking into account the requirements of both Ao I and transmission reliability in real-life scenarios.To capture the couplings between the age and transmission reliability of vehicular real-time status information,a novel utility function of vehicular network Qo S jointly designed based on the peak Ao I(PAo I)and transmission reliability is proposed,where the PAo I is defined as the peak values in the evolution of Ao I with time.Both accurate and approximate closedform expressions of average reliability,average PAo I and network Qo S utility are derived,and an adaptive vehicular status sampling algorithm is proposed to promote real-time and reliable vehicular state monitoring.Simulation results show that,compared with the conventional sampling algorithm,the proposed adaptive sampling algorithm can effectively improve the Qo S utility of vehicular networks on the basis of meeting the performance requirements of the monitoring service.Finally,considering the cyber-physical-social attributes of vehicles comprehensively,this dissertation investigates the timeliness of V2V/V2 I hybrid communication based information dissemination in vehicular social networks(VSNs),where the scale-free network theory is used to capture the characteristics of social relations among vehicles.Moreover,the network Ao I(NAo I),defined by the Ao I for the vehicle that lastly receives the information update originated from the base station(BS)in the VSN,is proposed to analyze the impact of both vehicular social relations and wireless communications on the timeliness of real-time vehicular information dissemination.Based on the novel framework of the mean-field theory,analytical results for real-time information dissemination in VSNs and the incurred average peak NAo I(PNAo I),i.e.,the time average of peak values in the evolution of the NAo I are derived.Furthermore,considering characteristics of wireless transmission links and social relations in VSNs,a novel scheme for joint optimization of the BS information update rate and vehicle transmit probabilities is proposed to improve the timeliness of real-time information dissemination.Simulation results show that compared with several baseline schemes,the proposed scheme can significantly reduce the average PNAo I of VSNs by up to 96%.