Cluster Based Intrusion Detection Technique for Wireless Networks

Abstract

Wireless networks are vulnerable to spoofing attacks, which allows for many other forms of attacks on the networks. Although th e identity of a node can be verified through cryptographic authentication, authentication is not always possible because it requires key management and additional infrastructural overhead. In this paper we propose a method for both detect ing spoofing attacks, as well as locating the positions of adversaries performing the attacks. We propose to use the spatial correlation of received signal strength (RSS) inherited from wireless nodes to detect the spoofing attacks. We then formulate the problem of determin ing the number of attackers as a multiclass detection problem. Cluster - based mechanisms are developed to determine the number of attackers. When the training data are available, we explore using the Support Vector Machines (SVM) method to further improve t he accuracy of determining the number of attackers. In addition, we developed an integrated detection and localization system that can localize the positions of multiple attackers. We evaluated our techniques through two test beds using both an 802.11 ( Wi - Fi ) network and an 802.15.4 network in two real office buildings. Our experimental results show that our proposed methods can achieve over 90 percent Hit Rate and Precision when determining the number of attackers. Our localizatio n results using a represen tative set of algorithms provide strong evidence of high accuracy of localizing multiple adversaries.