Research On Efficient And Reliable Data Transmission Mechanisms In Highly Dynamic Ad Hoc Networks

In recent years,with the rise of the Internet of Things(IoT),the construction of smart cities is advancing rapidly in various countries.In smart cities,ad hoc communications among highspeed(flying)vehicles can strengthen service function and expand service range,and provide better services for people.Therefore,highly dynamic ad hoc networks such as Flying Ad Hoc Networks(FANETs)and Vehicular Ad Hoc Networks(VANETs)have been widely used.To achieve stable and accurate data transmission is the core of deploying and applying highly dynamic ad hoc networks.How to design efficient and reliable data transmission mechanisms has become an urgent problem to be solved.FANETs and VANETs have their common characteristics,and there is a mutual reference of research between them.However,such highly dynamic ad hoc networks also have their own characteristics,which need to be taken into account when designing data transmission schemes.In FANETs,flying vehicles hover or move at high speed,and each node has limited communication energy with big data transmission demand.In VANETs,vehicle nodes move at high speed along roads and the change of network topology is predictable.Additionally,in those application scenarios where multiple services are provided in a highly dynamic ad hoc network simultaneously,the(flying)vehicles owning multiple sensors and network interface modules form a heterogeneous network.Good network scalability and effective Quality of Service(QoS)guarantee are essential.Considering the challenges above,this dissertation studies data transmission mechanisms in FANETs and VANETs from the following three aspects:"high-throughput and reliable data transmissions in FANETs","low-overhead and reliable data transmissions in VANETs”and "servicedifferentiation and reliable data transmissions in highly dynamic ad hoc networks(including FANETs and VANETs)".The main contributions and innovations of this dissertation include:1.To meet the transmission demand of a large amount of sensing data in FANETs,highthroughput and reliable data transmission mechanisms are studied,which improve data delivery rate and network throughput.In sensing applications such as environment monitoring where flying vehicles hover or move slowly,Radio Frequency(RF)links are stable,but with low capacity;Free-Space Optical(FSO)links have high capacity,but are vulnerable to disruption.Additionally,the uneven energy consumption and imbalanced traffic load cause part of network nodes to die quickly.Hence,the demand for large traffic transmission cannot be met.A Dynamic Energy&Traffic Balance(DETB)topology control mechanism based on RF/FSO adaptive switching,called RF/FSODETB,is proposed.Firstly,an adaptive RF/FSO switch mechanism is designed.According to weather conditions and distance between nodes,FSO links which may break in the near future will be switched to RF ones in advance.Secondly,a DETB topology control algorithm is designed under the premise that network link types(FSO or RF)are determined.It considers transmitting power,residual link bandwidth,and node energy simultaneously.Simulation results prove that RF/FSO-DETB can effectively reduce packet loss rate,extend network lifetime and improve network throughput.In the above mechanism,the type of each link needs to be determined according to the distance between nodes.Therefore,it is only applicable to quasi-static networks consisting of low-speed flying vehicles.In sensing applications such as visual battlefield communication in which flying vehicles move at high speed,highly dynamic topology and intermittent links result in high packet loss rate.Hence,the demand for large traffic transmission cannot be met.An Enhanced Network Coding-Aware Routing(ENCAR)is proposed,which executes path selection and data transmission based on the number of encoding nodes and route stability.If there are multiple routes whose route stability coefficients are greater than a preset threshold,a path with more encoding nodes will be selected.Simulation results show that ENCAR can significantly improve network throughput and packet delivery ratio.2.To reduce the huge overhead for topology maintenance and data transmission,lowoverhead and reliable data transmission mechanisms are studied,which reduce network overhead and improve packet delivery ratio.In application scenarios of sensing data collection,the VANET network topology keeps changing.A large amount of overhead for topology maintenance consume valuable network bandwidth resources,which affects reliable data transmissions.Considering that vehicles move along roads and drivers bear social attributes,the movements of vehicles are highly predictable A clustering m echanism based on model-driven predicted social pattern is proposed in this dissertation,called SAC(Social-Aware Clustering).Firstly,a discrete time-homogeneous semiMarkov model is built to describe each vehicle's movement,where the state transition probability and sojourn time probability distribution are inputs and each vehicle's social pattern is output.Secondly,a clustering algorithm is proposed,where nodes with the same social pattern(which may have similar driving routes in the future)are divided into the same cluster.Finally,a cluster head selection method based on the distance between vehicles,relative speed,and vehicle attributes is designed.Simulation results show that SAC can extend cluster lifetime and reduce network overhead,compared with traditional clustering mechanisms.In application scenarios in which delay-tolerant data transfer among long-distance vehicles,nodes "store,carry,forward" data.It is the key to data transmissions to choose the appropriate forwarding time and next hop node,which adds time dimension in the topology information.Additionally,it is difficult to guarantee the data delivery rate.In view of the fact that the trajectories of urban buses(not fully deterministic due to the influence of real-time traffic flow,road emergencies,etc.)have a high degree of regularity,this dissertation considers using a Delay Tolerant Networking-VANET(DTN-VANET)composed of bus nodes to transmit data for vehicle nodes.A routing mechanism based on data-driven predictive space-time graph is proposed,called PTSGR(Predicted Time-Space Graph Routing).Firstly,a frequent item mining algorithm based on time-space association rule is designed,which predicts the spatial and temporal information that bus nodes encounter,through mining history data rules.Secondly,a time-space graph model is constructed based on the predicted information,and then a sparse graph is generated by keeping links whose reliability exceed a given threshold.Finally,a routing algorithm based on the sparse graph is proposed.Simulation results show that PTSGR can reduce overhead and improve data transmission reliability.3.To guarantee different QoS in highly dynamic ad hoc networks(including FANETs and VANETs),service-differentiation and reliable data transmission mechanisms are studied,which provide QoS guarantee for various types of traffic.Considering the application scenario where delay-sensitive and reliable services are provided in a FANET under the Software Defined Networking(SDN)architecture,traffic flows of different services interact with each other,and the network topology and resources change dynamically.How to guarantee QoS for various types of traffic with limited network resources is a challenge.A Traffic-Differentiated Routing(TDR)mechanism is proposed in this dissertation.Firstly,a delay-and reliability-related cost function is defined.To obtain the perhop transmission reliability,a prediction model of link availability and node forwarding capability is proposed.Secondly,in the cost function,different weights are assigned to various flows according to their delay sensitivity and importance level and then a total weighted cost function is obtained.Finally,Nonlinear Integer Programming(NLIP)is employed to formulate mathematically the routing optimization problem.The objective is to minimize the total weighted cost of all traffic flows in the current time slot,while satisfying the typical constraints such as traffic conservation,maximum link capacity,and maximum end-to-end delay.An ant colony based heuristic algorithm is proposed to solve the optimization problem.The simulation results show that TDR can efffectively guarantee different QoS.The application scenario where delay sensitive and high-data-rate services are provided in a heterogeneous network with cellular network and VANET under the SDN architecture is also considered.On one hand,Multi-hop transmissions through the Dedicated Short Range Communication(DSRC)in VANET will result in long end-to-end delays for delay-sensitive traffic and QoS requirements cannot be satisfied;On the other hand,using the Long Term Evolution(LTE)link to carry high-rate data traffic can cause high cellular bandwidth cost.To achieve a balance between transmission delay and cellular bandwidth cost,a Traffic Differentiated Clustering Routing(TDCR)mechanism is designed.Firstly,a single-hop clustering strategy based on geographical locations is designed.In each cluster,the cluster head is selected according to the link duration and the distance to the nearby Road Side Unit(RSU),in order to realize data aggregation.Secondly,NLIP is employed to formulate mathematically the joint optimization problem of cluster head's delivery method(DSRC or LTE)selection and routing.The objective is to minimize the hop counts from the cluster head to the RSU under the typical constraints such as traffic conservation,maximum link capacity,and maximum endto-end delay.An optimal solution is obtained by solving the NLIP through CPLEX.A heuristic algorithm is also proposed to obtain a near optimal solution.The simulation results show that TDCR can effectively reduce the cellular bandwidth cost while guaranteeing the QoS.