A Fuzzy Trust Model for High-Density VANET Environments with Node Filtering

Vehicular Ad-hoc Networks (VANETs) in dense urban environments suffer from degraded routing reliability due to selfish or malicious vehicles disrupting cluster formation. This study introduces a fuzzy trust-based filtering mechanism to address the issue of untrustworthy Cluster Head (CH) selection. The approach dynamically combines direct and multilevel indirect trust using fuzzy transitivity and consensus models to prefilter low-trust nodes before clustering. Simulations using SUMO-generated mobility traces of 10,000 vehicles in Yogyakarta show that at a trust threshold Ï=0.5, the End-to-end Packet Forwarding Ratio (EPFR) improves by 24 under 60 selfish node conditions. At Ï=0.7, the Average Energy Consumption (AEC) for normal nodes is reduced by 25. Excessive filtering at Ï=0.9 decreases EPFR by 8. These results demonstrate that moderate thresholds provide the best balance between reliability and efficiency. This approach enhances VANET security and scalability without modifying existing clustering protocols. © 2025 IEEE.