Autonomous Driving Simulation Using Computer Vision

Abstract

This research presents a robust system for autonomous driving simulation utilizing Convolutional Neural Networks (CNNs) and OpenCV. The system demonstrates effective lane detection and obstacle avoidance mechanisms, achieving an accuracy of 90% in identifying road obstacles under simulated environments. By optimizing the CNN architecture, we achieved a 15% improvement in prediction accuracy compared to baseline models. (Note: This abstract is a summary. Please refer to the full PDF for detailed methodology and results.)