Date of Publication :12th November 2020

Abstract: Authentication of the genuine nodes is one of the important aspect in addressing the security issues in Vehicular Cloud Computing. The authentication methods followed involves the use of secret key generated by the credentials of the vehicle owner which is used for encryption or decryption of messages or to sign them. If this key gets compromised and malicious node gets access to this secret key then the complete service is disrupted. In order minimize the effect of compromised key we present a light weight continuous authentication scheme based on KNN classifier algorithm to validate the user, which can be installed along with the existing authentication schemes with little modifications. The simulation results show the proposed KNN classifier is the simplest and fastest classifier in comparison with other classifiers which makes the presented framework light weight. Using feature selection there is slight decrease in the accuracy but it does not degrade the performance of the framework to a greater extent. Security analysis shows that the proposed scheme is able to withstand most of the attacks efficiently in terms of Network Configuration Time, Kappa statistics and F- measure.

Reference :