BUDACVAB: Blockchain Utilized DriverAuthentication Considering Various Brands for VANETS

Vehicular Ad-Hoc Networks (VANETs) are specialized mobile networks that facilitate communication between vehicles and servers during driving. Through this communication, drivers connect to servers using various applications and share personal information with these applications. In VANET systems, au-thentication methods are used for security to access applications containing sensitive information on vehicles. The most common method for authentica-tion in VANETs involves analyzing the driving behavior of the driver using sensor data collected from the vehicle to verify the driver’s identity. This technique have some limitations which could stem from vehicle hardware, database capacity, or information security issues. Examples of limitations include the performance degradation of algorithms due to the usage of raw data, processing and capacity constraints of cars with low hardware, database capacity limitations, and the processing burden imposed by blockchain technology. Existing works does not consider authentication for a driver who uses a different brand car. This is a very common case for even a spouse who have different brand cars. In order to provide this property, various brands should provide data. This condition comes with a problem that different manufacturing companies should share data and agree on the conditions. This trusted environment can be provided by blockchain. In our study, we proposed a framework on VANET where different car manufacturers coexist in the same network in a trusted environment. Additionally , we enable drivers to securely access past driving data when they desire to use a car from a different brand. Blockchain has been integrated to establish a trusted environment for information exchange among different vehicle manufacturers. The integration of blockchain has created a constraint in recording raw driving data, which requires a large capacity. Therefore, in our study we applied Correlation-based Feature Selection (CFS) to the raw driving data collected at the end of the drive, resulting in the creation of a reduced dataset. Recording the reduced dataset on the blockchain for use as historical data in identity verification has decreased the capacity requirements on the blockchain and increased the Transaction per Second (TPS) value of the blockchain. The impact of CFS on identity verification algorithms was investigated, and it was observed that the performance of the algorithms improved. With this dataset, a %99 accuracy was achieved using the kNN algorithm. We also provided security protocols which provides secure communication between the entities in the framework. In our framework , blockchain is provided in between manufacturing peers of the brands. The RSU (Road Side Units) of VANETs are communicating with the manufacturing servers and these servers are communicating with the peers of the blockchain system. For this leveled structure, we provided security protocols in between different entities. Security proofs of these protocols are provided through AVISPA tool.