Roteamento baseado na informação espaço-temporal de trajetórias em redes veiculares Ad hoc

• Ano de defesa: 2013
• Autor(a) principal: Clayson Sandro Francisco de Sousa Celes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Temporal; Padrões de Trajetorias; VANET; Informação Espaço-; Roteamento; Roteamento (Administração de redes de computadores); Redes de computadores Administração; Computação
• Link de acesso: http://hdl.handle.net/1843/ESBF-AE7P4L

Vehicular ad hoc networks have frequently been used to offer solutions for data exchange in intelligent transport systems. Due to the characteristics of high mobility, frequent topology changes and rapid variations of scalability, one of the main challenges of VANETs is the design of robust schemes for message forwarding. In this sense, strategies based on geographic position have been often adopted by routing protocols. Moreover, the incorporation of in-vehicle navigation systems enabled an improvement in routing based on geographic information because the trajectories of vehicles shall be used for the routing process. However, there are several protocols that exploit the use of trajectories of vehicles, a few of them focus on the historical trajectories to aid in message routing process in the network. In this dissertation, routing schemes for vehicular networks, exploiting the use of information spatiotemporal trajectories of vehicles are presented. First, the focus is on the analysis of data collected from vehicles during their routine trajectories and the discovery of patterns of those trajectories, which denote routes that are often followed by users between a source and a destination. These patterns of trajectories are used to assist in the routing of messages in a VANET. The main contribution of this work is related to the increased rate of packet delivery among users belonging to sparse VANETs, which are seen as scenarios that make it difficult to build end-to-end routes. Besides the improvement of the delivery rate, simulation results show the efficiency of the proposed solution in detecting the best routes for routing messages, as well as a new approach to opportunistic routing based on the historical movement of vehicles.