Sistema de navegação para um robô móvel omnidirecional em competições de robótica

Abstract

This paper presents the development, evaluation, and documentation of software alternatives to equip an omnidirectional mobile robot, built with components from the LEGO Mindstorms EV3 kit, with autonomy and efficiency in performing tasks proposed by the RoboCup Brazil Challenge 2024, OPEN category. A distributed architecture was adopted using two EV3 controllers, and subsystems for omnidirectional movement based on mecanum wheels and PID control were implemented, along with odometric localization algorithms with correction and alignment strategies, internal graph representation for route planning via Dijkstra's algorithm, and a path control algorithm. Environmental perception was achieved using color, ultrasonic, and infrared sensors, incorporating signal processing strategies such as color classification using machine learning techniques. The results demonstrated that the autonomous system achieved satisfactory and competitive outcomes, securing 4th place in the OPEN category, despite variations and instabilities in more complex scenarios.