تشخيص گره مهاجم در شبكه هاي راديوي شناختي متمركز با استفاده از شبكه عصبي MLP

Detection of Malicious Node in Centralized Cognitive Radio Networks Based on MLP Neural Network

The cognitive radio network (CRNs) has been developed in recent years for the optimal use of Available vacuum in the frequency spectrum. In this network, Cooperate Spectrum Sensing (CSS) is used to combine the observations of all users. In CRNs, security is one of the most important problems. spectrum sensing data falsification (SSDF) attack is one of major challenges for CSS in CRNs, in which Malicious user are among honest users trying to change the information sent to the fusion center and thus make the fusion center wrong decision. In this paper, a method for defense against SSDF attack is proposed using MLP-based neural network. In this scheme, the weights of secondary users were constantly updated and finally the sensing results were combined in the fusion center based on their trusted weights. Simulation results show that the proposed scheme can significantly reduce the effects of Spectrum Sensing Data Falsification (SSDF) attack even percentage of malicious users are more than trusted users.

The cognitive radio network (CRNs) has been developed in recent years for the optimal use of Available vacuum in the frequency spectrum. In this network, Cooperate Spectrum Sensing (CSS) is used to combine the observations of all users. In CRNs, security is one of the most important problems. spectrum sensing data falsification (SSDF) attack is one of major challenges for CSS in CRNs, in which Malicious user are among honest users trying to change the information sent to the fusion center and thus make the fusion center wrong decision. In this paper, a method for defense against SSDF attack is proposed using MLP-based neural network. In this scheme, the weights of secondary users were constantly updated and finally the sensing results were combined in the fusion center based on their trusted weights. Simulation results show that the proposed scheme can significantly reduce the effects of Spectrum Sensing Data Falsification (SSDF) attack even percentage of malicious users are more than trusted users.