Navegação de veículos aéreos não tripulados para cobertura de áreas com minimização de tempo

• Ano de defesa: 2014
• Autor(a) principal: Gustavo Silva Castilho de Avellar
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Campos vetoriais; Cobertura; Veículos aéreos não tripulados; Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/BUBD-9NBGXR

This thesis addresses the development, implementation and testing of different methodologies for path planning and cooperation between Unmanned Aerial Vehicles (UAVs) in the area and border coverage tasks. Two di erent approaches were chosen and evaluated for this navigation problem. In the first one, each UAV is guided by implicit curves in a vector eld which is de ned in its workspace. The evaluation of the vector eld in each UAV position gives a vector which its direction is used as a reference for heading and its magnitude is used as a reference for speed. The second tested strategy is based on a intelligent waypoint distribution on a designated coverage area. With the location of the borders of the covered area and knowledge of the ight altitude and camera eld of view, the positions of the lines where the UAVs will y are calculated. This takes into consideration that the UAVs will cover the area with sweep like movements, doing turns outside the coverage region. The list containing the coordinates for the base and the waypoints that limit each sweep line, is used as an input to the second stage, as well as the number of available UAVs, their setup time, ight speed and maximum ight time. The set of points is modeled as vertexes in modi ed multiple salesmen traveling problem. The solution to this linear integer programming problem will be the number of UAVs that are needed to accomplish the mission in the minimum possible time and the sequence of waypoints that must be visited by each one of them, so that the area is completely covered. The contribution from this work is the realization that using all UAVs available for the mission will not always result in the minimum time to accomplish it. This is due to the fact that each UAV has a setup time, and in some situations, the usage of a bigger number of agents will not always make up for the additional setup time added by each one of them. Experiments were made with one and two UAVs covering the same region and validated these ndings.