Access Subnetwork Selection In Internet Of Vehicles

Internet of vehicles(IoV)is a newly emerging technology,which takes vehicles as communication carriers and is embedded with sensors,dedicated short range communication(DSRC),massive data processing and data integration technologies to provide all vehicles different kinds of network integrated services in the information network platform.IoV system is able to support vehicle to vehicle(V2V)communication,vehicle to infrastructure(V2I)communication and vehicle to pedestrian(V2P)communication.By taking the charateristics of IoV into consideration,such as fast changing network topology,severe fading wirless channel,limited channel bandwidth and preidictable vehicle trajectory,how to improve reliability of the transmission link,guarantee end-to-end queueing quality of service(Qo S)and choose a reasonable access sub-network for passengers are urgent problems in the IoV development.Therefore,this dissertation mainly investigates the following problems: link reliability analysis,network Qo S analysis,access network selection and scheduling strategy,through four aspects: outage probability,multi-hop end-to-end supermartingale queueing bound analysis,association strategy between passenger and trackside access point(TAP)and delay requirement based scheduling strategy.The main contributions of this dissertation can be summarized as follows.Given the problems of how to reduce the computation complexity and improve the real-time of the relay selection at the same time,the first part focuses on deriving the partial relay selection criterion in amplify-and-forward-based two-way relaying(TWR-AF)system.With the help of relaying vehicles in IoV,it is easy to accomplish wirelss communication between vehicles and access points(APs).Firstly,due to the severe diffraction during the communication among AP,mobile relay(MR)and vehicle,the wireless channel environment is very bad.Mixed Nakagami-m and ‘double' Nakagami-m fading channels are adopted to model the considered fixed-to-mobile(F2M)communication mode and mobileto-mobile(M2M)communication mode.Secondly,partial relay selection criterion is introduced to select the best relays according to partial channel information.A lower bound of system outage probability is first presented in concise closed form,which has a good approximation to the exact outage probability.By using this tight bound,we have derived the diversity and the coding gains as well as the average symbol error rate(SER)for the roadside mobile-relay-based TWR-AF.The optimum number of relay candidates is given in the form of a theorem.By using this theorem,one can easily determine how many relay candidates are required and sufficient for different sets of channel parameters,reducing the required complexity of the relay selection.Finally,simulation results highlight the theoretical analysis and show that the proposed analysis is useful for the design and performance evaluation of practical mobile-relay-based roadside TWR-AF systems.Given the problems of end-to-end queueing bound performance analysis of different types of data applications in the multi-hop IoV system,the second part focuses on deriving the tandem service process for multi-hop V2 V communication.In case that there is no direct communication link between the source node and the sink node,it is necessary to establish access subnetwork to deliver messages by cooperative V2 V communication.Firstly,the system model adopts ‘double' Nakagami-m fading to model the highly dynamic M2 M channel environment and considers the characteristics fo the stochastic arrived data application.Secondly,by utilizing SNC and martingale theory,the multi-hop end-to-end supermartgale backlog and delay bounds of delay sensitive application and delay tolerant application are analyzed under the static priority(SP),first in first out(FIFO)and earliest deadline first(EDF)scheduling policies.The impact on the network Qo S performance is investigated in multi-hop transmission IoV system.Finally,simulations are carried out by using the measured payload data of real wireless data traces from the circumstance of several subway lines in Seoul,Korea,which can roughly resemble the multi-hop IoV environment by adopting the Nakagami-m fading channel model,including voice over IP(Vo IP),gaming and UDP.The simulation results show that the supermartingale backlog and delay bounds can take full advantage of the stochastic characteristics of the data arrivals and services.The supermartingale end-to-end backlog and delay bounds are tight to the real data trace simulation results.End-to-end queueing performance can be regraded as criteria to decide the domain of the access subnetwork,which facilitates to provide valuable theoretical guidance for practical system design.Given the problems of whether the passengers should associate with the TAP or stay in the base station(BS),the third part focuses on deriving the association mechanism for the passenger and TAP in the IoV based heterogeneous wireless railway network.Due to the heavy data traffic flushing into the BSs of the cellular networks and the high data transmission rates of the TAP wireless link,TAPs act as a complement to the BSs in data delivery.Firstly,a symmetric game among the passengers is proposed to maximize their own payoffs.The formulation of the individual payoff model for each passenger consists of three elements,i.e.,players,reward and cost.The reward for the passenger is related to the achievable data rate and the cost is affected by the delay time they may encounter in the queue of the TAP.A payoff function is presented to facilitate the passenger to decide whether to associate with the TAP or not during the data transmission.Secondly,the system state information includes the queue length and the train position status.By applying the queueing and game theoretic-based approaches,the equilibrium strategies and the expected queueing delay performance of the passengers in the railway system are investigated under four different practical scenarios if the state information of the system are both known,partial known and both unknown.Finally,performance evaluations and discussions are also presented to illustrate the superiority of the proposed scheme when compared with max signal to noise ratio(SNR)scheme.Given the problems of different delay requirements for different data applications,the fourth part focuses on deriving the scheduling strategy by meeting the access subnetwork adaptation in the IoV based heterogeneous wireless railway network.Firstly,a service scheduling problem on minimizing the end-to-end delay of the different data services is formulated as an infinite-horizon time-average expected constraint Markov decision process(CMDP)problem.In order to provide the guidelines for scheduling selection scheme,a hybrid scheduling strategy is proposed to select an appropriate scheduling policy under different end-to-end delay requirements.Secondly,To obtain the theoretical end-to-end delay bounds of Vo IP and multimedia applications,the martingale theory is adopted to analyze the single-hop supermartingale delay bounds under the FIFO and EDF scheduling method.Finally,simulation results show that the delay performance of CMDP is much better than FIFO and EDF scheduling mechanisms.The hybrid scheduling strategy can adaptively select a suitable scheduling method to satisfy the delay requirement.From the simulation results,we can also see that the theoretical delay bounds are tight to the real data trace simulation results.