Analysis of the Application of Artificial Intelligence in Intelligent Transportation Systems
TL;DR
The study shows that AI technology has significantly improved traffic perception accuracy and real-time performance; however, it still needs to break through in the aspects of complex environment robustness, algorithm light weight, and the contradiction between privacy protection and data sharing.
Abstract
: Artificial intelligence provides core support for the efficient operation and safe control of intelligent transportation systems, which accelerates the urbanization process and makes people's lives more convenient. This paper reviews the application of deep learning-based target detection algorithms, multimodal sensor fusion, large-scale dataset construction, and other key technologies in intelligent transportation, analyzes the performance of target detection algorithms, such as You Only Look Once-X (YOLOX), Gaussian YOLOv3, and Complex-YOLO, in the detection of intelligent vehicles, explores nuScenes, Waymo Open Dataset (Waymo), CARRADA Dataset (CARRADA), and other large-scale datasets to support model training, and summarizes the problems that still exist in this field and the directions that can be developed in the future. The study shows that AI technology has significantly improved traffic perception accuracy and real-time performance; however, it still needs to break through in the aspects of complex environment robustness, algorithm light weight, and the contradiction between privacy protection and data sharing. This study provides a reference for the further development of intelligent transportation.