A Study on the Deep Learning Model for Autonomous Vehicle Driving

The purpose of this study was to investigate deep learning models for indoor autonomous vehicles. Deep learning was conducted using an NVIDIA model, transfer learning models, and the CNN model developed in this study, and the MAE and R2-score values were compared. The results demonstrated that both the NVIDIA model and the proposed CNN model exhibited not only stable loss values but also superior R2-score characteristics. By contrast, the transfer learning model did not achieve stable loss or R2-score values. Based on the results of the modified model, this study proposed a CNN model that is comparable to a GPU-based NVIDIA model.