Autonomous vehicles are becoming one of the developmental elements, not only for the transport of people but also in the field of data collection and monitoring, control of external elements or supervision and security. Their advantage is the ability to access dangerous areas which often cannot be accessed by humans. It is necessary that the vehicle recognizes its surrounding and reacts in an adequate way. In this work a study was carried out of the main techniques of artificial vision, machine learning and supervised learning applied in vehicles so they recognize the road and do not leave it. This work presents the viability of the different machine learning techniques for their application in the problem of autonomous driving. For this, an automobile robotic prototype has been constructed and an algorithm has been developed based on the Artificial Neural Network (ANN) algorithm and a user application which allows to carry out all integrated analysis and observe in real-time the vehicle’s view and the processing of the different snapshots. We have also demonstrated that the application of the stated algorithm, diverse processing techniques and artificial vision was sufficient, so that our robot could drive with precision and keep on the track of a road in a controlled environment.