False Information Mitigation Using Pattern-Based Anomaly Detection for Secure Vehicular Networks

Vehicular networks utilize wireless communication among vehicles and between vehicles and infrastructures. While vehicular networks offer a wide range of benefits, the security of these networks is critical for ensuring public safety. The transmission of false information by malicious nodes (vehicles) for selfish gain is a security issue in vehicular networks. Mitigating false information is essential to reduce the potential risks posed to public safety. Existing methods for false information detection in vehicular networks utilize various approaches, including machine learning, blockchain, trust scores, and statistical techniques. These methods often rely on past information about vehicles, historical data for training machine learning models, or coordination between vehicles without considering the trustworthiness of the vehicles. To address these limitations, we propose a technique for False Information Mitigation using Pattern-based Anomaly Detection (FIM-PAD). The novelty of FIM-PAD lies in using an unsupervised learning approach to learn the usual patterns between the direction of travel and speed of vehicles, considering the variations in vehicles’ speeds in different directions. FIM-PAD uses only real-time network characteristics to detect the malicious vehicles that do not conform to the identified usual patterns. The objective of FIM-PAD is to accurately detect false information in vehicular networks with minimal processing delays. Our performance evaluations in networks with high proportions of malicious nodes confirm that FIM-PAD on average offers a 38% lower data processing delay and at least 19% lower false positive rate compared to three other existing techniques.