Unlocking the Potential of VANETs: Trust-Based Authentication and Deep Learning for Enhanced Security and Efficiency

Abstract

Abstract Vehicular Adhoc Networks (VANETs) are emerging as a crucial component in the development of Intelligent Transportation Systems (ITS). These networks aim to enhance traffic operations, increase safety, and facilitate communication between vehicles and infrastructure. However, VANETs face significant privacy and security challenges. This study proposes a novel approach to address VANET energy efficiency and privacy analysis using a trust-based authentication system and deep learning methods. Attention layer integrated gradient kernel vector flow neural networks are utilized for classifying monitored data to detect malicious users. The hybrid multipath energy-efficient routing protocol contributes to improved energy efficiency. The experimental analysis evaluates energy efficiency, latency, throughput, packet delivery ratio, computational cost, and communication overhead, and trust value analysis. The proposed technique achieved energy efficiency of 99%, latency of 63%, throughput of 95%, PDR of 88%, computational cost of 57%, communication overhead of 59%, and trust value analysis of 77%. The experimental analysis reveals promising results in various performance metrics, demonstrating the potential of this approach in ensuring passenger and driver safety while addressing communication and security challenges in VANETs.