DETECTION AND ISOLATION OF BLACKHOLE ATTACK IN WIRELESS BROADBAND ECOSYSTEM USING ARTIFICIAL INTELLIGENCE

Subject Area: Electrical Engineering

This paper presents the detection and isolation of black hole attack in wireless broadband ecosystem using artificial intelligence technique. Many literatures were reviewed and research gap was established to justify the need for a new protection model. A case study 4G wireless network which used standard encryption algorithm was characterized using experimental methods and analysed with computer aided software engineering methodology to read blackhole attack. The threat was addressed via data collection, feature extraction, artificial neural network methods. These methods were designed using structural and mathematical approach and then implemented with SIMULINK and validated with tenfold cross validation approach. The result showed that with the neural network-based security algorithm, the average throughput achieved is 89.45% as against 56.94% without neural network. The percentage improvement of the throughput result achieved is 36.34%. The latency achieved with neural network is 88.10ms as against 186ms in the characterized network. The percentage improvement in the latency result is 47% which is very good.