| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 16.81 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A dissertação de mestrado apresentada insere-se no desenvolvimento do atual projeto europeu, Swarm of Biomimetic Underwater Vehicles. Este divide-se em duas etapas, tendo sido a primeira etapa representada pela construção de um veículo biomimético e a segunda pelo desenvolvimento de um cardume de veículos de baixo custo.
A segunda etapa do projeto SABUVIS, iniciou-se com a construção do primeiro protótipo de um desses veículos, denominado de Tobias. O foco principal desta etapa é a construção de um modelo de swarm, composto por diversos veículos, com capacidades operacionais idênticas ao protótipo já desenvolvido e de baixo custo. Este modelo tem como foco possibilitar a concretização de operações com uma tipologia benéfica para a Marinha Portuguesa, como a recolha de dados em missões do tipo reconhecimento e vigilância, a monitorização e reconhecimento de determinadas áreas ou a guerra anti-submarina.
O trabalho apresentado pretende realizar o desenvolvimento de um modelo de swarm, inicialmente com veículos de superfície, que futuramente possa ser inserido nos protótipos realizados. No modelo construído, são utilizados veículos de superfície diferenciais, isto é veículos apenas com dois veios e sem leme, devido à possibilidade de utilização de um maior número de veículos deste tipo e devido ao menor nível de complexidade perante veículos de sub-superfície.
A construção deste modelo é realizada em diversas etapas como, a preparação e adaptação de um simulador realista já existente, denominado de USVSim através do desenvolvimento de algoritmos de controlo de trajetória e seguimento, com a realização dos seus respetivos testes.
Após esta adaptação, foi concebido um sistema de controlo descentralizado, através de uma ferramenta de comunicações descentralizadas, o Multimaster, sendo os seus testes o controlo da formação dos veículos e a atribuição de tarefas especificas a cada um dos membros do grupo.
Pretendia-se implementar esses algoritmos num pequeno swarm de USVs, a construir para o efeito, mas devido a problemas na obtenção no material, tal não ix foi possível, sendo o trabalho desenvolvido inteiramente em simulação. Para tal, foi utilizado um simulador realista que opera em ambiente ROS. No entanto foi feito o levantamento do hardware necessário para construção futura desses veículos.
O modelo desenvolvido foca-se na coordenação e planeamento entre os indivíduos do grupo, permitindo a implementação em veículos com características distintas, permitindo um maior número de áreas de atuação.
The master’s thesis presented is part of the development of the current European project, Swarm of Biomimetic Underwater Vehicles. This is divided into two stages, the first stage is represented by the construction of a biomimetic vehicle, the BUV3 and the second by the development of a swarm of low cost vehicles. The second stage of the SABUVIS project started with the construction of the first prototype of one of these vehicles, called Tobias. The main focus of this stage is the construction of a low-cost model of a swarm, composed of several vehicles, with operational capabilities identical to the prototype already developed and at low cost. This model is focused on making it possible to carry out operations with a beneficial typology f or the Portuguese Navy, such as data collection, monitoring and reconnaissance of certain areas, anti-submarine warfare, or of a more scientific typology, such as hydrographic surveys or oceanographic studies. The work presented aims to develop a model of the swarm, initially with surface vehicles, which in the future can be inserted in the prototypes made. In the model built, differential surface vehicles are used, i.e. vehicles with only two shafts and no rudder, due to the possibility of using a greater number of vehicles of this type and due to the lower level of complexity compared to sub-surface vehicles. It was intended to implement these algorithms on a small swarm of USVs, to be built for this purpose, but due to problems in obtaining the material, this was not possible and the work was carried out entirely in simulation. To this end, a realistic simulator operating in the ROS environment was used. However, a survey was carried out of the hardware needed to build these vehicles in the future. The model developed focuses on coordination and planning between individuals in the group, allowing the characteristics of the vehicles where it is implemented not to be seen as a limiting factor.
The master’s thesis presented is part of the development of the current European project, Swarm of Biomimetic Underwater Vehicles. This is divided into two stages, the first stage is represented by the construction of a biomimetic vehicle, the BUV3 and the second by the development of a swarm of low cost vehicles. The second stage of the SABUVIS project started with the construction of the first prototype of one of these vehicles, called Tobias. The main focus of this stage is the construction of a low-cost model of a swarm, composed of several vehicles, with operational capabilities identical to the prototype already developed and at low cost. This model is focused on making it possible to carry out operations with a beneficial typology f or the Portuguese Navy, such as data collection, monitoring and reconnaissance of certain areas, anti-submarine warfare, or of a more scientific typology, such as hydrographic surveys or oceanographic studies. The work presented aims to develop a model of the swarm, initially with surface vehicles, which in the future can be inserted in the prototypes made. In the model built, differential surface vehicles are used, i.e. vehicles with only two shafts and no rudder, due to the possibility of using a greater number of vehicles of this type and due to the lower level of complexity compared to sub-surface vehicles. It was intended to implement these algorithms on a small swarm of USVs, to be built for this purpose, but due to problems in obtaining the material, this was not possible and the work was carried out entirely in simulation. To this end, a realistic simulator operating in the ROS environment was used. However, a survey was carried out of the hardware needed to build these vehicles in the future. The model developed focuses on coordination and planning between individuals in the group, allowing the characteristics of the vehicles where it is implemented not to be seen as a limiting factor.
Description
Keywords
Controlo Planeamento Coordenação Swarm