Detection of Malicious Nodes in WBAN using a Feed Forward Back Propagation Neural Network

Wireless Body Area Network (WBAN) is an emerging solution for local and distant health care, however the openness of wireless environment and the importance of people’s physiological data cause the exposure of this network to many attacks. Where, the attack of black-hole is among the most dangerous one. We have proposed in this paper a Feed Forward Back-Propagation Neural Network based method to detect malicious sensor nodes caused by the black hole attack in WBAN environment. For that, probabilistic features are extracted from each individual sensor node and a distance metric is calculated to classify sensor nodes. The WBAN performances in terms of delay, data rate and packets delivery ratio are calculated in order to measure and to evaluate the impact of illegitimate sensor nodes attacks. Moreover, the comparison of the proposed method to some recent similar state of the art methods shows its superiority in all the terms of evaluation metrics.