Generative AI Solution for Lane Departure, Pedestrian Detection and Paving of Autonomous Vehicle

This research article offers a thorough analysis of how Deep Learning (DL) might improve the performance and safety of Autonomous Driving (AD) systems. This research paper discusses the development and implementation of a custom object detection algorithm and real-time detection of lane markings in autonomous driving systems. The algorithm incorporates YOLOv8 (You Only Look Once) and ResNet-34 as its backbone, known for their effectiveness in detection tasks. The paper chronicles the training of this algorithms on a diverse dataset, consisting of images captured from unique angles, varied lighting conditions, and disparate road surfaces. Furthermore, it introduces a machine learning model dedicated to the classification of different types of lane markings, as soon as they are detected. The research provides an insightful discourse on the algorithm's design, training, and implementation, paving the way for enhanced safety measures in autonomous driving.