Diseño de una red neuronal para evitar colisiones de vehículos prototipo

Abstract

In 2022, Quito recorded 2,874 traffic accidents resulting in 176 fatalities and 2,348 injuries. This concerning situation stems from factors such as inadequate road infrastructure, insufficient signage, uncontrolled urban growth, and reckless behaviors among drivers and pedestrians. To address these challenges, a prototype intelligent vehicle has been developed. It is equipped with advanced hardware and software components, including high-performance microcontrollers and neural networks. This vehicle can anticipate potential collisions by detecting pedestrians, animals, and objects on the road. The system reacts dynamically: stopping or changing course upon detecting pedestrians or animals, and maneuvering around objects in its path. The central integration of sensors and a camera, along with YOLO algorithm-based object detection, enables real-time risk assessment and the implementation of proactive collision prevention measures. This research project aims to enhance urban safety through innovative technological solutions, marking a significant step towards mitigating traffic accidents on the road.