Controle de trajetória baseado em visão computacional utilizando o Framework ROS

Abstract

The goal of this work is to present the development of a path control for mobile robots based on computer vision, implemented in ROS (Robotic Operating System) Framework. The robot soccer environment was used as test platform for the proposed methodology. To accomplish this task was developed an algorithm to the robotic vision system, able to perform the calibration of image capture system and the identification of robots in the field according to the shape and color identification labels through the OpenCV library, which is integrated into the structure provided by ROS. The algorithms developed for computer vision, high-level control and low-level control were structured according to the guidelines of the ROS, therefore called nodes. The algorithms responsible for the high-level path control, image processing and control are performed on a desktop computer or notebook, while the algorithm for low-level control is running on a platforms Arduino embedded in the robots. The computer or notebook and embedded platform Arduino exchange informations among themselves in a distributed manner using topics, interconnecting nodes and transmit messages using the Publisher/Subscriber default. Also is addresses the development of a differential robot proposed, with their control devices, communications and traction.