Desenvolvimento de um sistema de navegação em ambientes internos para um robô pessoal

Abstract

This master thesis presents the development of a navigation system for a personal robot named MARIA in indoor environments. The navigation system enables the robot to move from one initial position and orientation to a final position and orientation within a previously mapped internal environment, while receiving voice instructions. The methodology involves: (i) generation of maps of the environment through the acquisition of sensory data from a Kinect sensor and a laser based proximity sensor, (ii) the position estimation robot of the upon the Monte Carlo localization algorithm (AMCL) (iii) path planning using a slightly modified wavefront and (iv) generation and execution of a collision free route. The two-dimensional map generated by the laser is used to estimate the position of the robot in the environment. The three-dimensional map generated using the Kinect is used to perform a two-dimensional projection of large obstacles found in the environment so it can be used for path planning. The methodology uses the idea that the most efficient way to navigate mobile manipulators is to use two-dimensional maps (2D) that take into account the height of the obstacles present in the path of the robot and the three-dimensional maps are used only when the robot needs to perform a task that requires of detecting and manipulating objects. To evaluate the methodology, the designed system was simulated and embedded in the robotic platform. The mobile manipulator responds to voice commands to perform the task, in addition to reacting in a satisfactory manner to the presence of unmapped objects in the environment.