We develop a self-driving car based on in-depth learning technique. This paper emphasizes to make it drive in the fixed lanes under lane line detection image processing and artificial neural network. The system will measure environment data, process them, and analyze technology in terms of image processing brought to detect object movements through applied streaming video image processing or .avi files from a webcam. This technology was manipulated for detecting lane lines. In the experiment, tracked roads while driving and tracked vehicle material movements through lane lines were more than 90%. Even though, the totally correct findings indicated that CNN system could recognize routes in the determined areas, analyze data, and make decision on its own correctly.

self-driving car, CNN system, ROS steering, simulation.