An Enhanced K Nearest Neighbor Classifier for Malicious Node Detection in VANET

Abstract

Recently, wireless communication technologies have become a vital part of our lives. The advancements made in communication technology, VANET systems is been introduced. With the increase of vehicles, different sorts of traffic are created in realistic environment. In some cases, the traffic is created by anomalies. Henceforth, the security of VANET communication becomes an important entity. In this paper, we proposed an enhanced k-Nearest Neighbor classifier that detected the malicious node in VANET. Generally, the classifier suffers from high computational cost in distance estimation during malicious node detection. The efficiency of the proposed classifier is experimented and implemented by validating the throughput and packet delivery rate. Compared to the existing classifier, the proposed classifier achieves 20-25% improvement. By doing so, the communication overhead and delay metrics have been achieved and also helps to minimize the computational storage costs.