Soft Computing Techniques for Intrusion Detection in Mobile Ad Hoc Networks

Abstract

Mobile ad hoc networks (MANETs) are suitable for communication in the absence of pre-defined infrastructure but it is extremely prone to threats due to its characteristics such as variable topologies, no centralized points followed by lack of a clear boundary of defense so that detection of intrusions is very difficult in MANETs than the conventional networks. The primary intention of intrusion detecting system is to categorize normal and abnormal behavior of nodes in the system. This paper focuses to develop local architecture of intrusion detection system using soft computing approaches for classifying the normal and abnormal activities. Results show that the proposed intrusion detection system is vital for detection of known and unknown attacks with high detection rates.