The Routing Protocols And Node Mobility Models Of Vehicular Ad Hoc Networks

With the development of wireless technologies, more and more vehicle based wireless devices and relative wireless communication standards were proposed. So the wireless communication enabled vehicles can communicate with each other or with roadside infrastructures. Such kind of network can be called Vehicular Ad Hoc Networks (VANET). Because of the potential benefit that VANET can bring to the Intelligent Transportation System (ITS), it has been developed to a hot research area. In this thesis, we focus on the application of Shanghai Traffic Information Grid and study the topics of the routing protocols and node mobility models of VANET based on the GPS data reported from over 4000 taxis in Shanghai City.In the first part of this thesis, we focus on the process of the taxi GPS data. Because of the interference of the high buildings in the metropolitan area, the GPS data itself is not accurate so that we have to determine the locations of taxis according to the GPS data which is called map-matching. Since the road network of Shanghai is very complex, the existing map-matching algorithms cannot work well. So we design a heuristic algorithm which considers the coordinate and direction of the taxi and the history map-matching results. The result shows that our map-matching algorithm obtains good accuracy. Meanwhile, taxis report their GPS data in a low frequency, averagely takes over 60 seconds to report a GPS data, which makes the absence of information. In order to make use of the data in the research of VANET, we have to interpolate the missed information to rebuild the continuous trace. So we design another heuristic algorithm which is based on the least turn rationale to determine the path that a taxi may travel during two consecutive GPS data report. Comparing to the shortest path algorithm, our solution is more efficient and more reasonable.Since current VANET researches are mostly based on simulation, the mobility model is critical to the research results. Most of the existing mobility models are based on probabilities and statistics. Routing protocols based on these mobility models are not suitable in the real VANET applications, which need to be modified according to the VANET characteristics. Besides, other mobility models which are extracted from real user data are often focus on the human mobility and also need modification. In this thesis, we propose a noval metropolitan taxi mobility model (META) which is extracted from the real taxi GPS data. We designed three parameters, which are called turn probability, road section speed and travel pattern, to capture the characteristics of running taxis. We compare the synthetic trace generated according to the META with the reconstructed trace on several metrics and the result shows that META can simulate the real taxi mobility accurately.The third part of this thesis is about the routing protocol of the VANET. We assume that the GPS device equipped taxis are also wireless communication enabled, so we need to design a routing protocol for the multi-hop data transmission service. Because of the high speed of taxis makes the network topology change fast, the conventional routing protocols of ad hoc networks, such as AODV, is not suitable anymore. So we use a new network paradigm, the Delay Tolerant Network (DTN), which allows the forward nodes to store the data until meet another node as next hop. Considering the constraint of wireless network bandwidth and node buffer size, existing DTN routing protocols cannot get satisfied performance. So we propose a new Distance Aware Epidemic Routing (DAER) protocol. By introducing the forwarding and buffer replacing strategies, DAER can reduce the consumption of network resources and increase the data delivery ratio. The simulation shows that DAER can obtain the expected performance.Finally, we focus on the application scenario of monitoring traffic status in urban area. We assume that vehicles are equipped with sensors and wireless devices which enable them to generate and transmit data about traffic status. Using the vehicle-to-vehicle and vehicle-to-roadside communication, traffic data can be transmitted to a certain sink node. In order to deal with the mismatch between the movement of taxis and the routing path of traffic data, we borrow the idea of semi-lunar valve of the artery in the human body and propose a mesh based routing protocol (MBRP). When the running direction of a taxi is reverse to the routing path of traffic data, the taxi should store the data on a nearby mesh node, and otherwise it should pick up data from the mesh node. The simulation results show that MBRP can obtain a high data delivery ratio.