O uso do algoritmo mapa de rotas probabilístico na robótica móvel: planejamento de caminhos para ambientes estáticos com foco em casas de vegetação

• Autor(a) principal: Magalhães Júnior, Antonio Mendes
• Data de Publicação: 2018
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/45521
• topic: Planejamento de caminhos; Robótica móvel; Casas de vegetação

Mobile Robotics is increasingly evolving in recent years; one of its main fields of research is the fundamental task of path planning, generating the shortest path for a robot towards the goal point in determinant environments. This study used the algorithm of Probabilistic Roadmaps (PRM) for path planning, considering the displacement of a simulated robotic platform of a greenhouse. The aim was to facilitate the applications of chemical agents for the control of agricultural pests in controlled environments. The algorithm received the starting and ending points of the robot, tracing the route to be traveled. It considered the smaller distance from the graph generated by the PRM and the constraints of the environment. For this, the representation of the environment, which is static and already known, was previously provided to the robot in off-line mode. Initially, the treatment of the images representing the maps was performed using the Matlab software in order to obtain image discretization. Different complexity maps were chosen in order to test the robustness and efficiency of the PRM algorithm. Using the path planning tool present in the Robotic System Toolbox of Matlab software, the PRM was run and a two-dimensional point set matrix was found that describes the least cost path to be covered. The simulated robot was then fed with this matrix and the robot parameters, such as size, linear velocity, and angular velocity were configured as required in the situation. The use of existing Matlab libraries favoredthe path planning and simulation taking into account the characteristics of the environment and the robot. Therefore, this study showed that the algorithm is robust and efficient for maps of different complexities, being a great tool option for path planning in static environments, which can be a promising method for greenhouse applications.