Research On Clustering And Hybrid Routing Mechanism Of Internet Of Vehicles

In recent years, urban traffic congestion and traffic safety problem has becomemore and more seriously, to improve the comfort and safety of transportation isimminent. Internet of Vehicles is an intelligent information service system, which is onthe basis of communication network, such as Vehicular Ad Hoc Networks (VANET),and uses the load on the vehicle electronic equipment to obtain road traffic information.Internet of Vehicles shares status information and real-time traffic information betweenmen and vehicles, vehicles and vehicles, vehicles and roadside fixed infrastructure aswell as vehicles and buildings.In order to meet the needs of intelligent transportation applications, and solve theproblems such as road congestion, traffic accidents effectively, dynamic routingtechnology is one of the key technologies of Internet of Vehicles and focused byscholars civil and abroad all the time. Internet of Vehicles features for the design ofefficient, reliable, and real-time routing protocol brings new challenges. These featuresinclude high-speed mobile vehicle nodes, frequent changes in network topology, unevendistribution of nodes, and the limited movement of vehicles. This paper has thoroughlystudied and analysed routing protocols need to meet the demand of the new designunder the Internet of Vehicles environment, proposes a new hybrid routing protocolTFBZRP (Traffic Flow-based ZRP) which is based on Zone Routing Protocol (ZRP).The main work and innovations are as follows:1) Carry on further research and analysis the performance of three classicalrouting protocols AODV、DSDV and ZRP under the Internet of Vehicles environment.The simulation result reveals that the performance of the ZRP protocol in multiplescenarios where vehicle nodes move with different speed and radius put up hugedifferences. With the limitations in driving speed, zone radius and mobile model, it hasno mean of improving the packet delivery rate, controlling overhead and reducing delay.ZRP cannot satisfy the design goal of Internet of Vehicles routing protocol.2) Aiming at the limitations in driving speed, zone radius, a new regional radiusadaptive algorithm based on traffic flow is presented, which adjusts the center nodeZone radius periodically by observing the change of Traffic flow. Optimum routingzone radius optis equal to the zone radius when IERP and IARP routing controloverhead ratio tend to be1for value (just like the scheme known as “Traffic Adaptive Estimation”).3) Regarding to the issues of unified period in NDM protocol, which makes theneighbor node status update not in time, a new NDM protocol named Traffic FlowBased Neighbor Discovery/Maintenance Protocol(TFBNDM) is presented, this is to sayHELLO beacon interval is aperiodic function of traffic flow Q. When Q increases,the HELLO beacon should increase the frequency of update to make shorter, on theother hand the HELLO beacon should reduce the frequency of update to make longer.4) Againsting the problem that redundant routing information increase the ZRPprotocol routing overhead, a new clusters algorithm named Biggest Zone Radius ClusterAlgorithm (BZRCA) is proposed, which establishes an applicable hybrid networkhierarchical structure. In this algorithm a node having biggest zone radius is selected asa primary center node, and the adjacent node whose status is not yet decided becomesthe member of the selected primary center node. When zone radius is the same, BZRCAuses the Lowest ID Cluster Algorithm (LIDCA) to select primary center node, and tagthe same radius value of the node as a secondary center node.5) Complete the module development and performance simulation of ZRPprotocol and TFBZRP protocol on VanetMobiSim and NS2co-simulation platform. Thesimulation results show that the TFBZRP increases packet delivery rate by28%, lowersthe routing overhead by80%and reduces the average end-to-end delay by44%at mostthan traditional ZRP. Therefore, TFBZRP can adapt to the Internet of Vehiclesenvironment where vehicle nodes have different speeds and topology changesfrequently. TFBZRP can increase computational complexity, however, and require moreCPU time and memory space.