Fishers Linear Discriminant Based Kullback–Leibler Divergence Based Attack Detection for Secured Communication in Manet

Abstract

A  Mobile ad hoc network (MANET) is an infrastructure-less network comprising the number of mobile nodes distributed randomly. In MANET, mobile nodes directly communicate with other mobile nodes within the radio communication range. Due to the random movements of the mobile nodes, the MANET suffers from several attacks. The various methods have been introduced for attack detection in MANET. But it failed to obtain a high-security level and less processing time. In order to improve data communication security, a Fisher's Linear Discriminant Based Kullback–Leibler Divergence Attack Detection (FLD-KLDAD) Mechanism is introduced with a higher detection rate and minimal processing time. Initially, the optimal features of a node such as cooperative count, residual energy and node mobility are calculated. After that, Fisher's linear discriminant technique classifies normal node and malicious node with the help of node features using a linear discriminant vector. After classification, the type of malicious attack is identified as an SYN flood attack or black hole attack using Kullback–Leibler Divergence (KLD). This helps to improve the detection rate and data communication security level. Simulation of FLD-KLDAD Mechanism and existing methods are carried out with different factors such as detection rate, processing time, data packet delivery ratio and end to end delay with respect to a number of nodes and data packets. The simulation results reported that the proposed FLD-KLDAD Mechanism obtains a high detection rate, data packet delivery ratio with minimum processing time as well as end to end delay.