An intelligent approach based on neuro-fuzzy detachment scheme for preventing jamming attack in wireless networks

Abstract

Securing information systems and the networks that connect them is paramount in the age of growing incidents of cyber attacks. Invasive and continuous cyber attacks on information systems continue to create a potentially shocking blow to the transmission of information from one end to the other. Wireless networks are often vulnerable to intentional interference physical layer reactive jamming attacks, in which the adversaries effectively and stealthily corrupt the packet by injecting high level of noise thereby keeping the channel busy. Consequently, legitimate traffic gets completely blocked; resulting in packet loss at the receiver side. These jammers cause serious threat to wireless communication and they are difficult to identify. A novel method known as ‘Neuro Fuzzy Detachment Scheme’ has been proposed to prevent such sophisticated and powerful jamming attack. The key idea is to identify and detach jammers from the network to ensure perspicuous transmission. The proposed detachment scheme operates in three modules. In module one, the past record log files are analyzed to extract the trust values for different traces. Using the extracted trust values, the decision factors are monitored for different runs with the attacker in between the communication range of source and the sink and then the suspicious traces are identified and classified using ANFIS classifier. In module two, to trace abnormality, three reorganizing algorithms have been developed to reorganize the categorized traces. In module three, misbehaving nodes are identified and detached from the network. It is observed through simulation studies that the proposed scheme attains higher throughput and packet delivery ratio while attaining lower delay.