Intelligent Vehicular Ad Hoc Networks which integrates deep learning techniques with modern vehicular communication networks can play a major role in prediction of vehicular traffic as well as efficient dissemination of critical information between vehicular nodes. An accurate traffic prediction mechanism is a key requirement for numerous applications of Intelligent Transportation System such as traffic management, accident prevention, route guidance and public safety. In this paper, a deep learning approach is proposed which is based on Convolutional Neural Network (CNN) combined with Temporal Convolutional Network (TCN), to predict the traffic patterns of vehicles. External factors like weather, weekend and holidays are considered along with internal factors such as location and time for analyzing their effect on vehicular traffic. Integration of CNN and TCN, captures spatio-temporal features, which are then merged with external factors to obtain a more accurate predicted traffic information. This predicted value is further disseminated within the vehicular network. Dataset of Indian cities is taken and converted to matrices of time vs space. Experimental results illustrate that our model outperforms other state-of-the-art techniques in regard to efficiency and accuracy.
Deep Learning Mechanism for Region Based Urban Traffic Flow Forecasting
Communic.Comp.Inf.Science
International Conference on Artificial Intelligence of Things ; 2023 ; Chandigarh, India March 30, 2023 - March 31, 2023
2023-12-03
15 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Vehicular flow forecasting , Intelligent Transportation Systems , Convolutional Neural Networks , Temporal Convolutional Network Computer Science , Artificial Intelligence , Information Systems and Communication Service , Machine Learning , Computer Applications , Computer Systems Organization and Communication Networks , Computer System Implementation
Traffic Flow Forecasting for Urban Work Zones
Online Contents | 2015
|Traffic Flow Forecasting for Urban Work Zones
IEEE | 2015
|A Spark-Based Deep Learning Method for Data-Driven Traffic Flow Forecasting
Europäisches Patentamt | 2020
|