A Secure VANET Model for Eavesdropping Attack Prevention

A Vehicular Ad Hoc Network (VANET) technology is becoming one of the most advanced technologies; however, the major barrier it faces is to keep the information secure and to improve its confidentiality. Recently, there have been some methods proposed to keep the information safe and implement the reliable platform for information transmission. However, based on the attacks varieties, and the way they are changing their approach, makes the system still insecure against some certain attacks. To overcome the vulnerability under the eavesdropping attack, we propose a solution to minimize the eavesdropping risk and protect the communication between vehicle and server. The solution can be considered as adding the road side unit (RSU). In fact, RSU is an extra neural network between vehicle and server which does not generate any key; however, it transfers the key from vehicle to the server. In this model, cipher text (encryption) is generated by RSU. In addition, the filters in the convolutional layers are used for encrypting the messages and to protect it against eavesdropping efficiently. We formally verify the security functionality of the solution scheme by training the neural networks based on different parameters. Analysis shows that our proposal has a strong ability to prevent eavesdropping attack.