This article describes the design and implementation of a 1:10 scale autonomous vehicle that combines computer vision techniques with classical control algorithms. The system is based on a distributed architecture where an ESP32-CAM camera captures images of the environment, which are processed on a computer using the YOLOv5 object detection model. The extracted information is used for real-time decision-making, such as stopping the vehicle in the presence of pedestrians or stop signs. The control signal is transmitted via WiFi to an ESP32-C3, which runs a PID controller to manage the vehicle’s actuators. The results obtained demonstrate that it is possible to implement low-cost, high-performance solutions for autonomous vehicle prototypes, with applications in both educational and research contexts.