Planejamento de sistema de direção autônoma em tempo real em um veículo com o auxílio de câmera

Abstract

This work presents the development of a vehicle capable of recognizing a path with the aid of a camera, planning the best trajectory, and navigating autonomously. The project utilizes ROS (Robot Operating System) as a framework for code development and Gazebo as a simulation environment, where a virtual track was created. Algorithms were implemented for track identification as well as the conversion of detected points into real-world coordinates. The vehicle control is based on the Model Predictive Control (MPC) approach, using the kinematic model of a two-wheeled differential robot. The expected results include metrics extracted from the vehicle’s odometry, such as average speed, maximum speed, total time, and the developed trajectory. This work serves as a foundation for the development of vehicles aimed at research in autonomous navigation.