Research On The Data Transmitting Schemes For Large-Scale VANET

As a special instance of Mobile ad hoc network (MANET) in the transportation field, Vehicular ad hoc network (VANET) can fuse different kinds of sensor data in road traffic environment and provide basic supports for the development of future Intelligent Transportation System (ITS). In the traditional road traffic system, the vehicles and the roads are relatively isolated since there is seldom effective information exchange between them. The VANET can reform this problem and facilitate the coordinated operation between the vehicles and the roads by vehicle-to-vehicle communications and vehicle-to-roadside communications. The VANET can be used to improve the traffic operation efficiency, vehicle safety and driving entertainment and it is with a broader development prospect. Recently, the VANET with its corresponding applications has aroused more and more attention from both the scientific research institutions and scholars, and it has been a hotspot in wireless network and ITS researches.In VANET, a reliable routing protocol is the basis for accuracy data delivery and also the prerequisite for implementing all kinds of applications. In the urban scenario, the large-scale VANET implies the network which covers vast geographic space and contains large quantity of vehicles. In large-scale VANET, the traditional routing protocols are no longer suitable for data transmitting, for the reason that characteristics of the large-scale VANET, such as massive amounts of nodes, highly partitioned topology and complex communication environment, make the traditional routing protocols high packet loss probability and overhead. Hence, designing a high performance routing protocol based on the large-scale VANET environment is with great significance for both the VANET and its applications.This thesis focuses on designing reliable data transmitting protocols for large-scale VANET. In order to simplify the complicated large-scale VANET, the concept of communication zone is proposed based on the modeling and forecasting of IVC (Inter-vehicle Communication) connectivity of the road section. Further, two different data transmitting protocols are proposed based on the communication zone structure respectively when there are RSUs in the network or not. The main work includes:1. In large-scale VANET, the road section is a basic unit for bearing the vehicle movements and the data transmitting links. This thesis firstly introduces the modeling and forecasting methods of IVC physical and logical connectivities of the road section. Specifically, a logical connectivity forecasting method based on data mining technology is presented. Taking the AODV protocol as an example, the simulated connectivity probabilities under different road traffic parameters are analyzed and the key factors which affect the logical connectivity of the road section are selected. Based on the simulated data, two logical connectivity forecasting models are presented, respectively the nonlinear regression-based model and the extreme learning machine (ELM)-based model. Simulation results show that the two models are both with high accuracy in predicting the logical network connectivity under different road traffic environments.2. In order to simplify the large-scale VANET in an urban scenario, this thesis proposes the concept of communication zone based on the macro distribution of IVC connectivity performance of road sections. Road sections which are with good performance of IVC connectivity are gathered together to form a virtual communication geographic area and establish the structure of the communication zone. The proposal of the concept provides a new thinking for the modeling of data transmitting issues in large-scale VANET. Besides, a communication zone partition scheme based on watershed algorithm is also presented.3. In order to solve problems such as high overhead and packet loss probability of the AODV protocol in large-scale VANET, an improved AODV protocol is presented based on restricted broadcasting by communication zones. After dividing the communication zones, a series of key communication zones which have good IVC connectivity performance are selected to construct a broadcasting geographic area. The route request broadcasting process of the AODV protocol is confined in the area so that the scale of the network is sharply decreased. Simulation results show that, compared with traditional AODV, the proposed improved protocol has advantages in packet delivery fraction and overhead.4. The distribution of vehicles in the road network is usually unbalanced. Hence, different RSU placement schemes can cause certain influence on the performance of the whole network. According to the road topology and vehicle distribution characteristics, this thesis proposes a maximum connectivity-based RSU placement scheme. The minimum coverage of each RSU is divided using the expansion and coloration algorihm. Further, the average connectivity model for all vehicles in the network is established. The RSU placement is formulated as a combinatorial optimization issue which is solved by intelligent algorithms. Compared with uniform placement and hot-spot placement, the proposed optimal placement scheme leads to the best connectivity.5. The communication zone partition issue is deeply analyzed when there are RSUs in the network. Further, this thesis proposes a new data transmitting scheme for large-scale VANET based on communication zones and RSU-assisted forwarding. In the scope of a communication zone, communications between vehicles and the RSU are achieved using the improved AODV protocol based on restricted broadcasting by the communication zone. Meanwhile, communications between RSUs are achieved by the backbone network. This scheme confines the wireless network into two communication zones which the source and the destination vehicles respectively locate in. Simulation results show that, compared with the AODV protocol and the improved AODV based on restricted broadcasting by communication zones, the proposed method can significantly improve the performance in packet delivery fraction and reduce overhead.