Self-localization and positioning vehicle navigation system based on computer vision and PID control

In this work a control system is presented for the autonomous navigation of vehicles using a monocular vision system to detect reference patterns in a controlled environment. This control system is based on the data extraction from the vision system camera feed to generate position vectors, which enable the localization of the vehicle in the environment, without having to depend purely on odometry avoiding the error that it generates. The infrastructure developed for this paper is a low cost 4-wheeled vehicle with a minimal quantity of sensors and a VGA camera for the computer vision system, all components commercially available.