In order for autonomous driving systems to function safely, object detection is essential for allowing vehicles to sense and comprehend their environment. Computer vision has undergone a revolution to deep learning techniques, which provide outstanding results in object detection tasks. This paper derives with a varied outline of object identification techniques based on deep learning designed for autonomous driving. It presents diverse deep learning architectures, datasets, and evaluation standard and throw a light on developments of YOLO variants. This paper presents a detailed research on the backbone Network used for object detection.