A genetic-Fuzzy Approach for Detection of Worm Attack in Ad-Hoc Wireless Networks

Abstract

Ad hoc networks are prone to internal and external attacks due to its complex properties such as dynamic topologies, no clear line of protection and resource constraints. Intrusion detection systems (IDS) has been designed to provide security to ad hoc network This paper proposed the genetic algorithm based feature selection to removes the irrelevant features and selects the relevant features from original dataset to improve the performance, reduced dimension and increased accuracy. This system is generally used to generate priceless solutions to optimization and search issues. Genetic algorithms which generate solutions to optimisation issues making use of procedures inspired by normal evolution, akin to inheritance, mutation and crossover Typically, a set of fuzzy rules (fuzzy classifiers) is used to define the behaviour in a computer network. Certain selected features are applied to the fuzzy classifier to detect the worm hole attack. A wormhole is a severe type of attack, where two adversaries are connected to each other through high speed off-channel link. In this, wormhole node receives the packet at one location and sends it to other wormhole node through high speed off-channel link. The main goal of this work is to detect the wormhole attacker node. The simulation is done using NS2, and the results of proposed genetic-based fuzzy logic intrusion detection system show good performance regarding detection rate, PDR, delay, throughput, false positive and false negative.