The burgeoning field of intelligent transportation systems(ITS)has been pivotal in addressing contemporary traffic challenges,significantly benefiting from the evolution of computational capabilities and sensor techno...The burgeoning field of intelligent transportation systems(ITS)has been pivotal in addressing contemporary traffic challenges,significantly benefiting from the evolution of computational capabilities and sensor technologies.This surge in technical advancement has paved the way for extensive reliance on deep-learning methodologies to exploit largescale traffic data.Such efforts are directed toward decoding the intricate spatiotemporal dynamics inherent in traffic prediction.This study delves into the realm of traffic prediction,encompassing time series,spatiotemporal,and origin-destination(OD)predictions,to dissect the nuances among various predictive methodologies.Through a meticulous examination,this paper highlights the efficacy of spatiotemporal coupling techniques in enhancing prediction accuracy.Furthermore,it scrutinizes the existing challenges and delineates open and new questions within the traffic prediction domain,thereby charting out prospective avenues for future research endeavors.展开更多
基金supported by the National Natural Science Foundation of China(62273057)。
文摘The burgeoning field of intelligent transportation systems(ITS)has been pivotal in addressing contemporary traffic challenges,significantly benefiting from the evolution of computational capabilities and sensor technologies.This surge in technical advancement has paved the way for extensive reliance on deep-learning methodologies to exploit largescale traffic data.Such efforts are directed toward decoding the intricate spatiotemporal dynamics inherent in traffic prediction.This study delves into the realm of traffic prediction,encompassing time series,spatiotemporal,and origin-destination(OD)predictions,to dissect the nuances among various predictive methodologies.Through a meticulous examination,this paper highlights the efficacy of spatiotemporal coupling techniques in enhancing prediction accuracy.Furthermore,it scrutinizes the existing challenges and delineates open and new questions within the traffic prediction domain,thereby charting out prospective avenues for future research endeavors.
