Restabelecimento da comunicação entre partições desconexas de uma rede de sensores sem fio utilizando veículos aéreos não tripulados

• Ano de defesa: 2015
• Autor(a) principal: Araujo, João Paulo de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de P os-Gradua ção em Ciência da Computa ção; UFLA; brasil; Departamento de Ciência da Computação
• Palavras-chave em Português: Rede de sensores sem fio; Veículo aéreo não tripulado (VANT); Rede tolerante a atrasos; Wireless sensor network; Unmanned aerial vehicle (UAV); Delay-tolerant network; Ciência da Computação
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/10644

One of the possible definitions for the lifetime of a Wireless Sensor Network (WSN) regards the time it keeps its connectivity. The formation of disjoint segments is related to physical damages to the nodes, internal node failure or energy depletion. Since disconnections compromise the operation of the network, providing alternatives to mitigate them can extend its lifetime. Several proposals tackle the problem of disconnections by means of redundant deployment or even using mobile nodes as relay stations. Differently from previous works, the present one proposes methods to federate disjoint segments of a WSN using Unmanned Aerial Vehicles (UAVs) as data mules, carrying physically packets among the segments. Thus, initially a cluster head is elected as the coordinator for each segment, since it is responsible for the interaction with the UAV when it comes to exchanging packets between segments. The routing process of messages inside a segment is performed by means of an extension of a geographical routing algorithm. And, in order to take the UAVs to the segments, three movement models are proposed. An initial model uses the minimum Hamiltonian cycle among the segments as the route for a single UAV. The second model is a traffic-aware one, since it takes into account the communication between segments to generate the route. Finally, the third one employs multiple UAVs to generate smaller routes that concentrate the traffic as well. Experiments were performed for each of the movement models in different scenarios. Results evince the effectiveness of the proposed solution, showing that each movement model has its performance correlated to the network traffic characteristics along with storage capacity of both sensor nodes and UAVs.