Accurate V2X traffic prediction with deep learning architectures.

Area of Science:

• Intelligent Transportation Systems
• Deep Learning for Traffic Prediction
• V2X Communication Networks

Background:

• Vehicle-to-Everything (V2X) communication offers potential for enhanced road safety and efficiency.
• Current V2X systems face challenges in data sharing and network reliability, hindering widespread adoption.
• Accurate traffic prediction is crucial for optimizing V2X functionalities.

Purpose of the Study:

• To propose and evaluate a novel Deep Learning (DL) approach for traffic prediction in V2X environments.
• To address challenges in data sharing and network reliability within V2X communication.
• To enhance the efficiency and safety of transportation systems through improved traffic forecasting.

Main Methods:

• Implementation of Bidirectional Long Short-Term Memory (BiLSTM) networks for traffic pattern prediction.
• Comparative analysis of BiLSTM against other Deep Learning architectures like unidirectional LSTM and Gated Recurrent Unit (GRU).
• Evaluation of prediction accuracy and performance metrics in simulated V2X scenarios.

Main Results:

• The Bidirectional Long Short-Term Memory (BiLSTM) model demonstrated superior accuracy in traffic pattern prediction compared to other DL architectures.
• The proposed DL approach effectively addresses data sharing and network reliability issues in V2X.
• Enhanced traffic prediction leads to more efficient resource allocation and improved network performance.

Conclusions:

• Deep Learning, specifically BiLSTM networks, offers a powerful solution for accurate traffic prediction in V2X communication.
• Improved traffic prediction capabilities contribute to enhanced road safety, reduced fuel consumption, and decreased emissions.
• The findings support the development of more sustainable and efficient intelligent transportation systems.