Traffic sign recognition system for autonomous vehicles

Smart vehicles, which have recently become popular in our daily lives, are vehicles that enable the driver to drive safely and comfortably by detecting the surrounding environment using the control units and sensors on them. The first of the three basic expectations from aoutonomous vehicles, which are the last stage in the transition from internal combustion vehicles to smart vehicles, is to receive the vehicle owner-passenger or control center requests correctly(Human-Machine interaction), the second is the correct management and administration of the vehicle, and the third is the road, traffic and objects in its surroundings determination-classification and estimation of its movements. While the first two expectations in autonomous vehicles are fully realized, there are still serious shortcomings in the third expectation. As a result of the studies in the field of artificial intelligence, many large companies in the automotive sector are working on autonomous vehicles. While these applications have enormous benefits for humans, the search for a better application always pushes scientists to improve existing studies. In one of these studies, the aim is to enable vehicles to interpret traffic signs and make decisions. In the study to be done, a convolutional neural network model that can classify the relevant traffic signs in different categories will be created. With this model, it is aimed to recognize the traffic signs by the autonomous vehicle by performing the classification of traffic signs, which is a very important task for autonomous vehicles. A model that recognizes the relevant traffic signs will be created and the programming language to be used is Python, convolutional neural networks (CNN) algorithm was chosen as the algorithm to be used.