Date Added: Nov 2012
This paper proposes the application of Adaptive Neuro-Fuzzy Inference System (ANFIS) in the design of a hybrid Wireless Intrusion Detection System (WIDS) that detects the black hole attacks in Mobile Ad-hoc NETworks (MANETs). Implementing security is a very challenging task in MANET due to its inherent vulnerabilities. The WIDS provides additional security to such networks and the proposed system consists of a hybrid scheme that combines the specification and anomaly based approaches. Executing the hybrid IDS using ANFIS improves the detection rate in MANETs. The proposed scheme uses Sugeno type FIS with Fuzzy C-Means clustering (FCM) and a hybrid neural network. The designed schema is tested for various network scenarios and the detection efficiency is found to be 96.7% as an average for all the cases.