A better way to detect sybil attacks in vehiuclar ad hoc networks

Sybil attacks, enabled by the anonymous nature of peer-to-peer broadcast communication in vehicular private networks (VANETs), pose a serious security threat. These attacks can significantly disrupt traffic flow, reduce efficiency, and potentially endanger traffic safety. Detecting Sybil attacks in VANETs is particularly challenging due to the dynamic network topology, real-time constraints, and decentralized nature of these networks. This paper proposes a novel Sybil attack detection method for VANETs, leveraging deep learning analysis of received signal strength indicator (RSSI) time series. The proposed system is designed to deliver effective results, even in brief interactions. Experimental results demonstrate the efficacy of our LSTM-based and CNN-based approaches, achieving 93.45% and 94.28% sensitivity in detecting attack messages, respectively.