Adaptive Modeling of Positional Falsification Malicious Behavior Using Type‐2 Fuzzy‐Based Aggregated Trust Analysis

Internet of Vehicles (IoV) involves a number of connected vehicles that can continuously share Basic Safety Messages (BSM). These messages contain details such as vehicle position, speed, and travel direction to improve traffic flow, but they also carry risks, including false positional attacks that can lead to unsafe driving decisions. Traditional methods in this domain primarily focus on privacy, identity verification, and fixed rules for detecting abnormal behavior. Due to this inability, they cannot fully capture precise vehicular motion. Also, it struggles to adapt to high motion vehicles, GPS noise, and hidden attack patterns. To address these limitations, this paper proposes TrustFRF (Trust‐aware Fuzzy Random Forest), an adaptive vehicular trust‐based detection model that continuously assesses vehicle trust based on its position consistency, motion patterns, and past‐time interactions, without relying on predefined attack patterns. The proposed model is evaluated using the publicly available “VANET malicious node dataset.” Experimental results suggest that TrustFRF achieves a notable detection accuracy of 99.40% with a low inference latency of 0.53 s for real‐time edge implementation.