A Novel Intrusion Detection System Based on Soft Computing Techniques Using Neuro- Fuzzy Classifier for Packet Dropping Attack in MANETs

Due to the advancement in communication technologies, mobile ad hoc networks are very attractive in terms of communication because mobile nodes can communicate without the relay on predefined infrastructure. Therefore, some complex properties of mobile ad hoc networks make it more vulnerable to internal and external attacks. From the security perspective, prevention based methods such as encryption and authentication are not considerably good solution for mobile ad hoc networks to eliminate the attacks so that intrusion detection systems are applied as a keystone in these types of networks. The main objective of intrusion detection system is to categories the normal and suspicious activities in the network. This paper proposed a novel intrusion detection system based on soft computing techniques for mobile ad hoc networks. The proposed system is based on neuro-fuzzy classifier in binary form to detect, one of vey possible attack, i.e. packet dropping attack in mobile ad hoc networks. Qualnet Simulator 6.1 and MatLab toolbox are used to visualize the proposed scenarios and evaluate the performance of proposed approach in mobile ad hoc networks. Simulation results show that the proposed soft computing based approach efficiently detect the packet dropping attack with high true positive rate and low false positive rate.