As the crucial component of intelligent transportation system, vehicular ad hoc networks (VANETs) are capable of providing a variety of safety-related functionalities and commercial-oriented applications, which significantly improves the driving experience. Due to the foreseen impact of VANETs, extensive studies in both academia and industry fields has been made, which emphasizes on effective VANETs implementations. In practical VANETs scenarios with open wireless communication characteristics, enhanced security strategies should be deployed in order to guarantee transmission safety. Moreover, individual vehicle needs to perform pre-defined authentication process toward all the acquired messages, some of which may be generated by abnormal devices or malicious attackers. In this case, with large amounts of anomaly messages to be authenticated during a relatively short time period, the denial of service (DoS) attack is possible. Note that the vehicle has limited computation capability and restrained storage. In this paper, we address the above issues by developing a secure and efficient authentication scheme with unsupervised anomaly detection. In our design, certificateless authentication technique is deployed for conditional privacy preserving, along with the Chinese remainder theorem for efficient group key distribution and dynamic updating. Subsequently, the corresponding unsupervised anomaly detection method is illustrated, which applies dynamic time warping for distance measurement. The proposed method could remarkably alleviate unnecessary authentication burden in vehicle side. DoS attack can also be prevented in this way. Furthermore, anomaly detection method is conducted by the involving road side units (RSUs), while the contents of the processing traffic flows are kept secret to RSUs during the entire process. Security analysis shows that our scheme can achieve desired security properties. Additionally, performance analysis demonstrates that our design is efficient compared with state-of-the-art.
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Materials Science(all)