masterThesis
Um sistema de visão para navegação robusta de uma plataforma robótica semi-autônoma
Fecha
2006-05-19Registro en:
BEZERRA, João Paulo de Araújo. Um sistema de visão para navegação robusta de uma plataforma robótica semi-autônoma. 2006. 83 f. Dissertação (Mestrado em Automação e Sistemas; Engenharia de Computação; Telecomunicações) - Universidade Federal do Rio Grande do Norte, Natal, 2006.
Autor
Bezerra, João Paulo de Araújo
Resumen
Large efforts have been maden by the scientific community on tasks involving locomotion of mobile robots. To execute this kind of task, we must develop to the robot the ability of navigation through the environment in a safe way, that is, without collisions with the objects. In order to perform this, it is necessary to implement strategies that makes possible to detect obstacles. In this work, we deal with this problem by proposing a system that is able to collect sensory information and to estimate the possibility for obstacles to occur in the mobile robot path. Stereo cameras positioned in parallel to each other in a structure coupled to the robot are employed as the main sensory device, making possible the generation of a disparity map. Code optimizations and a strategy for data reduction and abstraction are applied to the images,
resulting in a substantial gain in the execution time. This makes possible to the high level decision processes to execute obstacle deviation in real time. This system can be employed in situations where the robot is remotely operated, as well as in situations where it depends only on itself to generate trajectories (the autonomous case)