HYBRID ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM AND WHALE OPTIMIZATION ALGORITHM FOR VANET SECURITY ENHANCEMENT

Abstract

The increasing reliance on connected vehicles and real-time data exchange in Intelligent Transportation Systems (ITS), security concerns in Vehicular Ad Hoc Networks (VANETs) have become more pressing due to their dynamic and decentralized nature. This study presents a robust and adaptive intrusion detection framework that integrates Adaptive Neuro-Fuzzy Inference System (ANFIS) with Whale Optimization Algorithm (WOA) to enhance security in VANET environments. The approach begins by preprocessing the NSL-KDD dataset through redundancy removal, label encoding, and Min-Max normalization. Dimensionality is reduced using Random Projection, and optimal features are selected through the Enhanced Chimp Optimization Algorithm (EChOA). These selected features are then fed into ANFIS, with its parameters fine-tuned by WOA to improve classification accuracy. Results demonstrate the model's effectiveness, achieving 99.6% accuracy, 99.8% precision, 99.71% F1-score, and 99.57% recall, outperforming existing VANET security methods. The findings highlight the potential of combining neuro-fuzzy systems with bio-inspired optimization for reliable and efficient intrusion detection in vehicular networks.