Throughput analysis in vehicle-to-vehicle networks

In recent years, providing wireless high-speed data connection service to fast moving vehicles on highways has been the focus of research in wireless technology. One approach is to use vehicle-to-vehicle multihop communication for packet relay, known as Vehicular Ad-hoc Networks (VANET). In this thesis, we propose a framework that supports high data rate connection using vehicle-to-vehicle multihop communication. We propose a generic TDMA-based MAC layer protocol, with base stations acting as coordinating points. Multihop relaying is used to connect source and destination nodes when they are out of transmission range of each other. Base stations select intermediate relay nodes for each connection, and assign channels to these nodes.;One key parameter in VANET is the network throughput, in terms of number of requests being accepted by the base station. Wireless signal interference is incurred when an antenna receives signals from various sources in the same channel. Therefore, with the limited number of channels, relays should be chosen, and channels should be assigned such that interference is avoided, and maximum number of requests is accepted. We design and implement greedy algorithms for channel assignment for two types of problems, online scheduling and offline scheduling. We identify four network parameters that affect network throughput: number of channels, transmission range, node density and request intensity. We then evaluate this algorithm against a derived theoretical average under steady network state (free flow) using different values of the network parameters. Our results show that in steady network state, the network throughput under greedy algorithm is about 70 percent of the theoretical average throughput. Offline scheduling algorithm outperforms online scheduling algorithm in all steady state test cases. Simulation results on steady network state also demonstrate that when inter-node spacing is comparable to transmission range, change of node density has significant impact on network throughput.;We also tested the greedy channel assignment algorithm using common transient states in VANET. We find out that this algorithm is resistant to slight network disturbances such as stop-and-go waves, but is seriously affected by scenarios such as gaps in the traffic flow.