Research On Unicast Routing Mechanisms For Vehicular Ad Hoc Networks In City Scenarios

In recent years, there is a growing interest on the research and deployment of inter-vehicle communication system. Inter-vehcle communication can bring great convenience to people. For example, when traveling on the road, people can get the traffic information through the road communication system and adjust their travel line timely depending on the traffic information. Furthermore, transportation system can also adaptively adjust the traffic lights by traffic flow and traffic information to enhance efficiency in Intelligent Transportation Systems(ITS). Be seen from the above, the inter-vehicle wireless communications technology can greatly improve traffic safety and improve traffic flow conditions. In inter-vehicle communications, mobile ad hoc network technology has many advantages, such as low latency, network robustness and network topology.VANET(Vehicular Ad Hoc Networks) is a specific type and one of the most promising applications of mobile ad hoc networks, which is formed by vehicles that are equipped with wireless communication devices. As Vehicular Ad Hoc networks is self-organization without network infrastructure or central control node and high dynamic topology due to high speed, how to design an efficient routing protocols to delivery packets opportunely and reliablely has been a hot topic for Vehicular Ad Hoc Network.In this dissertation, the author's major research work in this area is summarized as follows:A novel geography and traffic based routing algorithm for VANETs (Vehicular Ad hoc Networks) in city environments was proposed. A contention-based forwarding routing algorithm for VANETs in city environments was proposed. This dissertation proposed an analysis of the contention-based forwarding (CBF) routing algorithm for VANETs in city environments and the lower bound of the timer was given on the condition that the suppression scheme accomplished successfully at give probability. This dissertation gave the formula for End-to-End delay in spare Vehicular Ad hoc Networks, and an algorithm named DEA(Delay Evaluation Algorithm) was proposed to evaluate the End-to-End delay in routing algorithm. In this dissertation, the probabilities of Vehicular Ad hoc Network connectivity for difference cases were derived. Based on the probabilities, a service node placement strategy for Vehicular Ad hoc Networks(VANETs) in city scenarios was proposed. And the service node can be auxiliary nodes in routing algorithm or Internet Access Gateway. A simulation platform is constructed to verify the correctness and to analyze the performance of the above algorithms.(1) Geography and Traffic Based Routing for Vehicular Ad Hoc Networks in City Scenarios. A novel geography and traffic based routing algorithm for VANETs (Vehicular Ad hoc Networks) in city environments was proposed, which using real time traffic information and mobility prediction to forward data with the idea of "carry and forward". Furthermore, this protocol can work well where the destination node is moving. Simulation result shows that it achieves significant delivery rate improvement and lower delay compared other similar routing approaches.(2) A Contention-Based Forwarding Routing Algorithm. A contention-based forwarding routing algorithm for VANETs in city environments was proposed, which using distributed contention-based strategy to select the next hop based on the positions of the neighbor nodes without HELLO messages. The novel scheme forwards packets along an anchor-based path which is determined according real time traffic information and the topology of the network, with the idea of "carry and forward" to deal with partitioned networks. Simulation result shows that it achieves significant delivery rate improvement and lower delay compared other similar routing approaches. This dissertation proposed an analysis of the contention-based forwarding (CBF) routing algorithm for VANETs in city environments, which is very important to determine the value of the timer used in CBF. And the analysis is one beneficial and important component of CBF. The lower bound of the timer was given on the condition that the suppression scheme accomplished successfully at give probability. The analysis is validated by comparing simulation and analysis results. (3) A Benchmark for Delay Performance in Spare Vehicular Ad hoc Networks. In spare networks, node forwards packet using "carry and forward" to improve the delivery rate. This dissertation gave the formula for End-to-End delay in spare Vehicular Ad hoc Networks, and an algorithm named DEA(Delay Evaluation Algorithm) was proposed to evaluate the End-to-End delay in routing protocol. DEA aims to minimize End-to-End delay. Taking the neighbor relationship of nodes as input, DEA outputs the minimal End-to-End delay and the routing path with the minimum delay. DEA can also be used in a network routing where nodes move regularly.(4) On the Connectivity of One-Dimensional Vehicular Ad Hoc Networks. In this dissertation we analyzed connectivity of one-dimensional Vehicular Ad Hoc Networks where vehicle gap distribution can be approximated by an exponential distribution. The probabilities of Vehicular Ad hoc Network connectivity for difference cases are derived:(1) The probability of connectivity of the ith vehicle and the jth vehicle(j>i). (2) The probability of connectivity of the vehicles in the interval [0,L](L>0) (3) The probability of connectivity of a fixed node at point zero and the geographical point L(L>0). The analysis is validated by computing the probability of network connectivity and comparing it with the Mont Carlo simulation results.(5) Connectivity-based Service Node Placement in Vehicular Ad Hoc Networks. This dissertation proposed a service node placement strategy for Vehicular Ad hoc Networks in city scenarios. Firstly based on the probability of connectivity and the characteristic of VANETs, the model of optimal placement of the service nodes was described. Furthermore, one solution based on PSO(particle swarm optimization) for this optimal problem was given. Finally, we validated our scheme by simulation.