The current computational advance allows the development of technological solutions to everyday problems, for example, today mobile robots are used for the automated transport of raw materials. Therefore, the study of applications such as obstacle avoidance and autonomous navigation with mobile robots are of great importance in academia, driving research that seeks to contribute to scalable and low-cost solutions with mobile robots. Next, a proposal to use artificial vision and add classification functionalities to the navigation control of a mobile robot is described. The solution consists of integrating an embedded system into the mobile robot's navigation control which, through a classification model, identifies the closest traffic signal and then sends a command to the mobile robot so that it modifies its navigation. This work seeks to verify the technical feasibility of using electronic development boards to develop Machine Learning solutions, by obtaining a classifier with an accuracy of 94.9% and a loss rate of 0.21.