Real-Time Detection of Malicious Nodes Using Poplar Optimized Sparsity-Aware Network for Improving VANET Security

Vehicular Ad Hoc Networks (VANETs) constitute a fundamental part of the Intelligent Transportation Systems since they allow vehicle to vehicle, and vehicle to infrastructure communications. However, due to the nature of Mobile Ad hoc Networks and high mobility of VANETs, they are vulnerable to malicious node attacks which adversely affect the reliability of the network and road safety. Current detection frameworks are not optimized for real-time and can be computationally expensive, along with having high false positives. In this work Real-Time Detection of Malicious Nodes using Poplar optimized Sparsity-Aware Network for Improving VANET Security, a real-time malicious node identification using Deep Sparsity-Aware Neural Network (DS-ANN) with Reformed Poplar Optimization (RePO) is performed. The goal is to improve detection performance and minimize computations to achieve better scalability while being suitable for various VANET environments. The extensive simulation reveals that the proposed RePO-DS-ANN model can detect undesired remnants with a detection accuracy of 99.80 % while maintaining a false-positive rate of under 0.20 % and a latency that is over 40 % less than the existing methods. It also has good throughput when the offered traffic rates and attack models are changed. Therefore, RePO-DS-ANN is a novel advancement in VANET security that can provide real-time vehicular environment with scalable, efficient, and enhanced accurate malicious node detection.