Research On Physical Topology Properties Of Vehicular Ad Hoc Networks In Highway Scenarios

T RANSPORTATION systems are playing a critical role in virtually all facets of modern life; however, despite such progress, increasing mobility is creating undesirably negative impacts on economy and individual quality of life. More than 43,600 people were killed and another 3.5 million were injured on US highways in the year of 2006. The cost to the economy of all motor vehicle crashes is approximately $230.6 billion, occupying 2.3% of the US GDP. Furthermore,3.7 billion hours of traveling were delayed and 2.3 billion gallons of fuel were wasted in the 85 selected US urban areas.Conventional approaches are not suitable for tackling problems associated with transportation systems, such as traffic accidents, congestions, and environment. An alternative approach refers to intelligent transportation systems (ITS). As an integral part of ITS, vehicular ad hoc networks (VANET) which are a specific form of mobile ad hoc networks (MANET) and wireless sensor networks (WSN), can provide direct or multi-hop vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-pedestrian (V2P) communications based on pre-existing road layouts. The emerging and promising VANET technology, which has drawn tremendous attention from government, academics, and industry in the past several years, has also been envisioned as one of the forefront research hotspots and increasingly available for a large number of cutting-edge applications as diverse as cooperative collision warnings, intersection decision supports, adaptive traffic lights and scheduling, real-time traffic information disseminations, and interactive multimedia services such as peer-to-peer file sharing, video transmission, online gaming as well as high-speed Internet access.Good progress has recently been made in the fields of dedicated medium access control (MAC), data dissemination, securing communications, and mobility modeling; however, there are few studies concerning the topological properties of vehicular ad hoc networks. Due to the unique characteristics of VANET, e.g., temporal and spatial distribution of vehicles, high mobility with an organized but constrained pattern, and a variety of radio propagation conditions, the corresponding graph structure will rapidly change, which in turn leads to frequent network fragmentation, thus the wireless links between any two nodes cannot be guaranteed all the time. What is more, the quasi-linear topology may reduce the possible link redundancy. For this reason, we focus on what we called the fundamental properties of vehicular ad hoc networks, e.g., degree distribution, clustering coefficient, average path length, and the connectivity. Major research work includes:(1) We describe in Chapter 1 the concept of VANET, which has been made viable by the convergence of micro-electro-mechanical systems (MEMS), mobile computing, and wireless communications, and then indicate the similarities and differences with MANETs and WSNs. The communication architecture for VANET is outlined, the protocol stacks are investigated, and the potential vehicular networks applications are explored. The VANET-related consortia, standards, and projects around the globe are presented. The enabling technologies, such as spectrum allocation, media access, and data dissemination, for the realization of vehicular networks are also reviewed.(2) We introduce in Chapter 2 the three types of models for the study of VANET: vehicular mobility model, radio propagation model, and network topology model. We also describe our modeling approaches based on the analysis of each specific model.(3) Based on the distribution of headway under free flow conditions in the traffic flow theory, we derive in Chapter 3 the probability of nodes which have at least k neighbors within the radio coverage, for better understanding the intrinsically analytic expression between network size n, radio communication range r, and the probability pk of nodes with at least k neighbors. Besides, we also derive the relationship between average node degree<k> and macroscopic traffic density k, based on the percolation theory and scene segmentation technology.(4) The clustering coefficient and average path length are discussed in Chapter 4 and 5, respectively. We obtain the empirical expectation and probability distribution of each network parameter through simulations.(5) The connectivity of vehicular ad hoc networks, which is a prerequisite to providing reliable applications to the network users, is fully studied in Chapter 6. We derive a probability analysis algorithm to calculate the necessary but not sufficient condition of k-connected VANET in highway scenarios, namely the minimum node degree. We as well obtain the sufficient and necessary condition to ensure that the network is 1-connected, namely the giant component size.Extensive experiments were undertaken to verify the derived results, which are of practical value for researchers and developers who perform design and implementation of vehicular ad hoc networks.