| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 29.51 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Ao longo dos últimos anos, o Laboratório de Sistemas Autónomos (LSA) sediado no Instituto Superior de Engenharia do Porto (ISEP), tem vindo a participar em diversos projetos nacionais e europeus, nas múltiplas áreas da robótica móvel. Fruto deste trabalho faz agora parte de um projeto europeu, cujo propósito é o desenvolvimento de um sistema de mineração subaquático capaz de operar em minas inundadas a céu aberto, no qual ´e responsável pelo desenvolvimento dos sistemas de localização e mapeamento. No âmbito desta dissertação foi desenvolvido o sistema responsável por realizar o mapeamento em tempo real de todo o ambiente subaquático, requerido para a operação do veículo de mineração. Todavia, de modo a possibilitar a implementação do mesmo, várias camadas de software responsáveis por parametrizar, filtrar, reconstruir e visualizar, foram desenvolvidas. Com base numa análise da literatura, o sensor escolhido para a implementação deste sistema foi o sonar multi-feixe. Contudo, tal como as câmaras, também os sonares necessitam de ser calibrados e parametrizados, sendo em muitos casos, necessária uma parametrização dinâmica e constante do sensor. Foi este problema que levou ao desenvolvimento do sistema capaz de parametrizar os sonares automaticamente, definindo parâmetros como alcances e ganhos. Com este objetivo, foram utilizados não só a informação do próprio sensor, mas também informações externas, como dados da navegação do veículo a qual os sonares estavam acoplados. Paralelamente ao sistema de parametrização, foi também desenvolvido o sistema de mapeamento e visualização, responsável por reconstruir o mundo em tempo real. Os resultados obtidos demonstraram que o sistema foi capaz de realizar os objetivos propostos, tendo sido o mesmo implementado e testado na demonstração final do projeto no qual esta dissertação se enquadra, o VAMOS.
Over the last few years, the LSA based in ISEP has been participating in several national and European projects in the multiple mobile robotics areas. As result of this work, the research laboratory is now participating in a European project whose purpose is the development of an underwater mining system capable of operating in open-pipe mines, in which is responsible for the development of the mapping and localization systems. In the scope of this dissertation, a system capable of making a real time 3D mapping of the the entire underwater environment required for the operation of the mining vehicle, was developed. However, in order to enable its implementation, several layers of software responsible for tuning, filtering, reconstruction and visualization have been developed. Based on a studied literature, the chosen sensor for the implementation of this system was the multibeam sonar. However, just like the cameras, the sonars also need to be calibrated and parameterized, and in many cases, dynamic and constant parameterization of the sensor is required. Was this problem that led to the development of the multibeam sonar automatic tuning system, defining parameters as ranges and gains. For this purpose, not only information from the sensor itself, but also external information such as navigation data of the vehicle which the sonars were coupled, was used. Parallel to the parameterization system, the mapping and visualization system was also developed, responsible for the real time 3D reconstruction of the environment. The results obtained demonstrated that the system was able to accomplish the proposed objectives, having been implemented and tested in the Viable Alternative Mine Operating System (VAMOS) project final demonstration.
Over the last few years, the LSA based in ISEP has been participating in several national and European projects in the multiple mobile robotics areas. As result of this work, the research laboratory is now participating in a European project whose purpose is the development of an underwater mining system capable of operating in open-pipe mines, in which is responsible for the development of the mapping and localization systems. In the scope of this dissertation, a system capable of making a real time 3D mapping of the the entire underwater environment required for the operation of the mining vehicle, was developed. However, in order to enable its implementation, several layers of software responsible for tuning, filtering, reconstruction and visualization have been developed. Based on a studied literature, the chosen sensor for the implementation of this system was the multibeam sonar. However, just like the cameras, the sonars also need to be calibrated and parameterized, and in many cases, dynamic and constant parameterization of the sensor is required. Was this problem that led to the development of the multibeam sonar automatic tuning system, defining parameters as ranges and gains. For this purpose, not only information from the sensor itself, but also external information such as navigation data of the vehicle which the sonars were coupled, was used. Parallel to the parameterization system, the mapping and visualization system was also developed, responsible for the real time 3D reconstruction of the environment. The results obtained demonstrated that the system was able to accomplish the proposed objectives, having been implemented and tested in the Viable Alternative Mine Operating System (VAMOS) project final demonstration.
Description
Keywords
Perceção subaquática Sonar multibeam Parametrização automática Tempo real Reconstrução tridimensiona VAMOS ROS Underwater perception Multibeam sonar Automatic tuning Real time 3D reconstruction