Um sistema óptico de reconhecimento de trajetórias para veículos automáticos.

Abstract

This dissertation presents a trajectory recognition optical system for automated vehicles, whose main characteristics are low cost, simplicity and high sampling rate. The main application for this system is in the recognition of nonreflective optical guides painted on even floors, for guiding automated vehicles. The detection of the reference trajectory is done by the Image Acquisition Unit (IAU), which acquires low-resolution images from the floor. These images are quantized and arranged i n a vector. The image vector is processed and trajectory error variables are estimated by the Image Processing Unit (IPU). A trajectory controller architecture for a prototype vehicle is also presented. It consists of the Trajectory Control Unit (TCU), which commands the vehicle drive system to make i t track the nonreflective optical guide with a constant speed. The components of this system were evaluated using a test platform, simulation or by experiments with the prototype vehicle. The experimental results with the prototype vehicle demonstrated the satisfatory functioning of the proposed system.