Detecting Malicious Roadside Units in Vehicular Social Networks for Information Service

The accurate identification of malicious nodes in vehicular social networks (VSN) can ensure the secure and efficient operation of the mobile network. The roadside unit (RSU) undertakes various tasks in the internet of vehicles (IoV) and processes a large amount of data. It plays an indispensable and crucial role in the IoV. Therefore, the damage and scope of the attack on RSU are more significant. This paper proposes a trust model for detecting malicious RSUs (TMDMR) to realize the trust decision problem of the roadside unit in VSN. The trust management model is constructed by designing two main decision indicators (QoS trust and social trust) in vehicular social networks, and the algorithm is intended to complete the calculation of global trust. Finally, the global trust server implements the trust decision of RSU. The simulation results demonstrate that TMDMR is better than the comparison scheme in terms of Precision and Recall. When the malicious number of nodes is 45% during OA attack, the Precision of TMDMR is 5.3% and 26.2% higher than comparison schemes, and Recall value is also 4.7% and 30.1% higher than them, respectively. When the malicious number is 45%, TMDMR has the lowest packet dropping rate (23.8% and 44.7%) to the comparison schemes. Its end-to-end delay is 42% and 51.3% lower than other two schemes. It also has advantages in terms of response time to complete a round of detection.