Fuzzy trust model for trustworthiness of information in vehicular ad hoc network

• Main Author: Soleymani, Seyed Ahmad
• Format: Thesis
• Language: English
• Published: 2019
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/96280/1/SeyedAhmadSoleymaniPSC2019.pdf.pdf; http://eprints.utm.my/id/eprint/96280/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:143032
• Institution: Universiti Teknologi Malaysia

Vehicular ad hoc networks (VANETs) represent a class of ad hoc networks created to enhance road safety, passenger comfort, traffic efficiency, and reduce overall traffic accidents. In this network, all applications are based on the exchange of data among vehicles, hence, the trustworthiness of data and vehicles is essential. The presence of selfish nodes, as well as obstacles, by generating inaccurate and incomplete information, has a negative impact on the trustworthiness of the vehicular environment. Therefore, the aim of this research is to propose a trust model in a vehicular environment, which results in the safety and comfort of passengers, by increasing the trustworthiness of information. For this purpose, a fuzzy trust model (F-TRUST) composed of three modules, namely, plausibility, experience, and decision-making, was proposed. To cope with the inaccurate and incomplete data, the proposed model evaluated the trust level of both data and vehicles by performing fuzzy logic in both line-of-sight (LOS) and non-line-of-sight (NLOS) conditions. The proposed model was evaluated by well-known evaluation measures such as precision, recall, F-measure, overall accuracy, and communication overhead. The results indicate that F-TRUST had better performance as compared to the weighted voting (WV) approach. In addition, the F-TRUST scheme outperformed the WV approach under various patterns of attacks such as simple attack, opinion tampering attack, and cunning attack. In conclusion, this study demonstrates that F-TRUST can improve the trustworthiness of information objectively, and in turn help vehicles to detect the selfish nodes and inaccurate data.