Abstract:

Vehicular Ad hoc Networks (VANETs) are built on intelligent cars and may support vehicle-to-vehicle (V2V) and vehicle-toroadside unit (V2R) communications. In this study, we offer a model for forecasting network traffic by taking into account the characteristics that might cause road traffic to occur. The suggested model incorporates a Random ForestGated Recurrent Unit-Network Traffic Prediction algorithm (RF-GRU-NTP) to estimate network traffic flow based on road and network traffic at the same time. This model is divided into three phases: network traffic prediction based on V2R communication, road traffic prediction based on V2V communication, and network traffic prediction taking into account both V2V and V2R communication. The proposed hybrid model, which is implemented in the third phase, selects the important features from the combined dataset (including V2V and V2R communications) using the Random Forest (RF) machine learning algorithm, and then applies deep learning algorithms to predict network traffic flow, with the Gated Recurrent Unit (GRU) algorithm providing the best results. The simulation results reveal that the proposed RF-GRU-NTP model outperforms previous network traffic prediction algorithms in terms of execution time and prediction errors.

Description:

Vehicular network, network traffic prediction, road traffic prediction, regression methods, classification methods, machine learning algorithms, deep learning algorithms