Short-Term Traffic Prediction With Deep Neural Networks: A Survey

Abstract

In modern transportation systems, an enormous amount of traffic data is generated every day. This has led to rapid progress in short-term traffic prediction (STTP), in which deep learning methods have recently been applied. In traffic networks with complex spatiotemporal relationships, deep neural networks (DNNs) often perform well because they are capable of automatically extracting the most important features and patterns. In this study, we survey recent STTP studies applying deep networks from four perspectives. 1) We summarize input data representation methods according to the number and type of spatial and temporal dependencies involved. 2) We briefly explain a wide range of DNN techniques from the earliest networks, including Restricted Boltzmann Machines, to the most recent, including graph-based and meta-learning networks. 3) We summarize previous STTP studies in terms of the type of DNN techniques, application area, dataset and code availability, and the type of the represented spatiotemporal dependencies. 4) We compile public traffic datasets that are popular and can be used as the standard benchmarks. Finally, we suggest challenging issues and possible future research directions in STTP.INDEX TERMS Artificial intelligence, deep neural network (DNN), intelligent transportation systems (ITS), neural networks, prediction algorithms, short-term traffic prediction (STTP), traffic forecasting

A comprehensive study of speed prediction in transportation system: From vehicle to traffic

“…This section discusses the deep learning methods in terms of spatial dependency, temporal dependency, spatio-temporal dependency, and external factors modeling. Based on the generational classification of deep-learning methods for traffic prediction in ( Lee et al., 2021 ), this article does not focus on the first generation methods (i.e., Deep Belief Networks and Stacked AutoEncoder (SAE)) but pays attention to the latest deep learning methods.…”

“…However, those researches paid insufficient attention to recent advances, e.g., attention and graph-based learning models. The deep-learning methods were categorized into five generations to describe the research trend in ( Lee et al., 2021 ), and reference ( Ye et al., 2020 ) provided a survey specifically on graph-based deep learning. Meanwhile, research ( Yin et al., 2021a ) conducted experiments to compare different deep learning methods, and research ( López Manibardo et al, 2021 ) discussed the pros and cons of deep-learning methods in detail.…”

Semantic Based Moving Weighted Average Model for Road Traffic Prediction

“…[ Mahmuda Akhtar, et.al (2021)] analysed various existing models based on machine learning methodologies and discussed the strength and weaknesses of the models. In recent research in traffic prediction systems [K. Lee, et al (2021)] shows massive data production. Nowadays, deep learning methodologies are applied for traffic prediction.…”

Traffic prediction with deep learning for vehicular communication network

“…In modern society, due to the continuous expansion of the automobile market and the increase of traffic demand, people all over the world are easy to face the problem of road congestion, which is not conducive to the productivity of modern society and seriously affects people's life comfort. As an inevitable trend, the Internet of vehicles will exert a positive impact on the effective intelligent transportation management and intelligent travel management of cities [1]. As an important part of the construction of vehicle networking, traffic flow prediction has made great progress.…”